1 @c Copyright (C) 1988,1989,1992,1993,1994,1995,1996,1997,1998,1999,2000,2001,2002
2 @c Free Software Foundation, Inc.
3 @c This is part of the GCC manual.
4 @c For copying conditions, see the file gcc.texi.
7 @chapter Target Description Macros and Functions
8 @cindex machine description macros
9 @cindex target description macros
10 @cindex macros, target description
11 @cindex @file{tm.h} macros
13 In addition to the file @file{@var{machine}.md}, a machine description
14 includes a C header file conventionally given the name
15 @file{@var{machine}.h} and a C source file named @file{@var{machine}.c}.
16 The header file defines numerous macros that convey the information
17 about the target machine that does not fit into the scheme of the
18 @file{.md} file. The file @file{tm.h} should be a link to
19 @file{@var{machine}.h}. The header file @file{config.h} includes
20 @file{tm.h} and most compiler source files include @file{config.h}. The
21 source file defines a variable @code{targetm}, which is a structure
22 containing pointers to functions and data relating to the target
23 machine. @file{@var{machine}.c} should also contain their definitions,
24 if they are not defined elsewhere in GCC, and other functions called
25 through the macros defined in the @file{.h} file.
28 * Target Structure:: The @code{targetm} variable.
29 * Driver:: Controlling how the driver runs the compilation passes.
30 * Run-time Target:: Defining @samp{-m} options like @option{-m68000} and @option{-m68020}.
31 * Per-Function Data:: Defining data structures for per-function information.
32 * Storage Layout:: Defining sizes and alignments of data.
33 * Type Layout:: Defining sizes and properties of basic user data types.
34 * Escape Sequences:: Defining the value of target character escape sequences
35 * Registers:: Naming and describing the hardware registers.
36 * Register Classes:: Defining the classes of hardware registers.
37 * Stack and Calling:: Defining which way the stack grows and by how much.
38 * Varargs:: Defining the varargs macros.
39 * Trampolines:: Code set up at run time to enter a nested function.
40 * Library Calls:: Controlling how library routines are implicitly called.
41 * Addressing Modes:: Defining addressing modes valid for memory operands.
42 * Condition Code:: Defining how insns update the condition code.
43 * Costs:: Defining relative costs of different operations.
44 * Scheduling:: Adjusting the behavior of the instruction scheduler.
45 * Sections:: Dividing storage into text, data, and other sections.
46 * PIC:: Macros for position independent code.
47 * Assembler Format:: Defining how to write insns and pseudo-ops to output.
48 * Debugging Info:: Defining the format of debugging output.
49 * Cross-compilation:: Handling floating point for cross-compilers.
50 * Mode Switching:: Insertion of mode-switching instructions.
51 * Target Attributes:: Defining target-specific uses of @code{__attribute__}.
52 * Misc:: Everything else.
55 @node Target Structure
56 @section The Global @code{targetm} Variable
58 @cindex target functions
60 @deftypevar {struct gcc_target} targetm
61 The target @file{.c} file must define the global @code{targetm} variable
62 which contains pointers to functions and data relating to the target
63 machine. The variable is declared in @file{target.h};
64 @file{target-def.h} defines the macro @code{TARGET_INITIALIZER} which is
65 used to initialize the variable, and macros for the default initializers
66 for elements of the structure. The @file{.c} file should override those
67 macros for which the default definition is inappropriate. For example:
70 #include "target-def.h"
72 /* @r{Initialize the GCC target structure.} */
74 #undef TARGET_COMP_TYPE_ATTRIBUTES
75 #define TARGET_COMP_TYPE_ATTRIBUTES @var{machine}_comp_type_attributes
77 struct gcc_target targetm = TARGET_INITIALIZER;
81 Where a macro should be defined in the @file{.c} file in this manner to
82 form part of the @code{targetm} structure, it is documented below as a
83 ``Target Hook'' with a prototype. Many macros will change in future
84 from being defined in the @file{.h} file to being part of the
85 @code{targetm} structure.
88 @section Controlling the Compilation Driver, @file{gcc}
90 @cindex controlling the compilation driver
92 @c prevent bad page break with this line
93 You can control the compilation driver.
96 @findex SWITCH_TAKES_ARG
97 @item SWITCH_TAKES_ARG (@var{char})
98 A C expression which determines whether the option @option{-@var{char}}
99 takes arguments. The value should be the number of arguments that
100 option takes--zero, for many options.
102 By default, this macro is defined as
103 @code{DEFAULT_SWITCH_TAKES_ARG}, which handles the standard options
104 properly. You need not define @code{SWITCH_TAKES_ARG} unless you
105 wish to add additional options which take arguments. Any redefinition
106 should call @code{DEFAULT_SWITCH_TAKES_ARG} and then check for
109 @findex WORD_SWITCH_TAKES_ARG
110 @item WORD_SWITCH_TAKES_ARG (@var{name})
111 A C expression which determines whether the option @option{-@var{name}}
112 takes arguments. The value should be the number of arguments that
113 option takes--zero, for many options. This macro rather than
114 @code{SWITCH_TAKES_ARG} is used for multi-character option names.
116 By default, this macro is defined as
117 @code{DEFAULT_WORD_SWITCH_TAKES_ARG}, which handles the standard options
118 properly. You need not define @code{WORD_SWITCH_TAKES_ARG} unless you
119 wish to add additional options which take arguments. Any redefinition
120 should call @code{DEFAULT_WORD_SWITCH_TAKES_ARG} and then check for
123 @findex SWITCH_CURTAILS_COMPILATION
124 @item SWITCH_CURTAILS_COMPILATION (@var{char})
125 A C expression which determines whether the option @option{-@var{char}}
126 stops compilation before the generation of an executable. The value is
127 boolean, nonzero if the option does stop an executable from being
128 generated, zero otherwise.
130 By default, this macro is defined as
131 @code{DEFAULT_SWITCH_CURTAILS_COMPILATION}, which handles the standard
132 options properly. You need not define
133 @code{SWITCH_CURTAILS_COMPILATION} unless you wish to add additional
134 options which affect the generation of an executable. Any redefinition
135 should call @code{DEFAULT_SWITCH_CURTAILS_COMPILATION} and then check
136 for additional options.
138 @findex SWITCHES_NEED_SPACES
139 @item SWITCHES_NEED_SPACES
140 A string-valued C expression which enumerates the options for which
141 the linker needs a space between the option and its argument.
143 If this macro is not defined, the default value is @code{""}.
145 @findex TARGET_OPTION_TRANSLATE_TABLE
146 @item TARGET_OPTION_TRANSLATE_TABLE
147 If defined, a list of pairs of strings, the first of which is a
148 potential command line target to the @file{gcc} driver program, and the
149 second of which is a space-separated (tabs and other whitespace are not
150 supported) list of options with which to replace the first option. The
151 target defining this list is responsible for assuring that the results
152 are valid. Replacement options may not be the @code{--opt} style, they
153 must be the @code{-opt} style. It is the intention of this macro to
154 provide a mechanism for substitution that affects the multilibs chosen,
155 such as one option that enables many options, some of which select
156 multilibs. Example nonsensical definition, where @code{-malt-abi},
157 @code{-EB}, and @code{-mspoo} cause different multilibs to be chosen:
160 #define TARGET_OPTION_TRANSLATE_TABLE \
161 @{ "-fast", "-march=fast-foo -malt-abi -I/usr/fast-foo" @}, \
162 @{ "-compat", "-EB -malign=4 -mspoo" @}
167 A C string constant that tells the GCC driver program options to
168 pass to CPP@. It can also specify how to translate options you
169 give to GCC into options for GCC to pass to the CPP@.
171 Do not define this macro if it does not need to do anything.
173 @findex CPLUSPLUS_CPP_SPEC
174 @item CPLUSPLUS_CPP_SPEC
175 This macro is just like @code{CPP_SPEC}, but is used for C++, rather
176 than C@. If you do not define this macro, then the value of
177 @code{CPP_SPEC} (if any) will be used instead.
179 @findex NO_BUILTIN_SIZE_TYPE
180 @item NO_BUILTIN_SIZE_TYPE
181 If this macro is defined, the preprocessor will not define the built-in macro
182 @code{__SIZE_TYPE__}. The macro @code{__SIZE_TYPE__} must then be defined
183 by @code{CPP_SPEC} instead.
185 This should be defined if @code{SIZE_TYPE} depends on target dependent flags
186 which are not accessible to the preprocessor. Otherwise, it should not
189 @findex NO_BUILTIN_PTRDIFF_TYPE
190 @item NO_BUILTIN_PTRDIFF_TYPE
191 If this macro is defined, the preprocessor will not define the built-in macro
192 @code{__PTRDIFF_TYPE__}. The macro @code{__PTRDIFF_TYPE__} must then be
193 defined by @code{CPP_SPEC} instead.
195 This should be defined if @code{PTRDIFF_TYPE} depends on target dependent flags
196 which are not accessible to the preprocessor. Otherwise, it should not
199 @findex NO_BUILTIN_WCHAR_TYPE
200 @item NO_BUILTIN_WCHAR_TYPE
201 If this macro is defined, the preprocessor will not define the built-in macro
202 @code{__WCHAR_TYPE__}. The macro @code{__WCHAR_TYPE__} must then be
203 defined by @code{CPP_SPEC} instead.
205 This should be defined if @code{WCHAR_TYPE} depends on target dependent flags
206 which are not accessible to the preprocessor. Otherwise, it should not
209 @findex NO_BUILTIN_WINT_TYPE
210 @item NO_BUILTIN_WINT_TYPE
211 If this macro is defined, the preprocessor will not define the built-in macro
212 @code{__WINT_TYPE__}. The macro @code{__WINT_TYPE__} must then be
213 defined by @code{CPP_SPEC} instead.
215 This should be defined if @code{WINT_TYPE} depends on target dependent flags
216 which are not accessible to the preprocessor. Otherwise, it should not
219 @findex SIGNED_CHAR_SPEC
220 @item SIGNED_CHAR_SPEC
221 A C string constant that tells the GCC driver program options to
222 pass to CPP@. By default, this macro is defined to pass the option
223 @option{-D__CHAR_UNSIGNED__} to CPP if @code{char} will be treated as
224 @code{unsigned char} by @code{cc1}.
226 Do not define this macro unless you need to override the default
231 A C string constant that tells the GCC driver program options to
232 pass to @code{cc1}, @code{cc1plus}, @code{f771}, and the other language
234 It can also specify how to translate options you give to GCC into options
235 for GCC to pass to front ends.
237 Do not define this macro if it does not need to do anything.
241 A C string constant that tells the GCC driver program options to
242 pass to @code{cc1plus}. It can also specify how to translate options you
243 give to GCC into options for GCC to pass to the @code{cc1plus}.
245 Do not define this macro if it does not need to do anything.
246 Note that everything defined in CC1_SPEC is already passed to
247 @code{cc1plus} so there is no need to duplicate the contents of
248 CC1_SPEC in CC1PLUS_SPEC@.
252 A C string constant that tells the GCC driver program options to
253 pass to the assembler. It can also specify how to translate options
254 you give to GCC into options for GCC to pass to the assembler.
255 See the file @file{sun3.h} for an example of this.
257 Do not define this macro if it does not need to do anything.
259 @findex ASM_FINAL_SPEC
261 A C string constant that tells the GCC driver program how to
262 run any programs which cleanup after the normal assembler.
263 Normally, this is not needed. See the file @file{mips.h} for
266 Do not define this macro if it does not need to do anything.
270 A C string constant that tells the GCC driver program options to
271 pass to the linker. It can also specify how to translate options you
272 give to GCC into options for GCC to pass to the linker.
274 Do not define this macro if it does not need to do anything.
278 Another C string constant used much like @code{LINK_SPEC}. The difference
279 between the two is that @code{LIB_SPEC} is used at the end of the
280 command given to the linker.
282 If this macro is not defined, a default is provided that
283 loads the standard C library from the usual place. See @file{gcc.c}.
287 Another C string constant that tells the GCC driver program
288 how and when to place a reference to @file{libgcc.a} into the
289 linker command line. This constant is placed both before and after
290 the value of @code{LIB_SPEC}.
292 If this macro is not defined, the GCC driver provides a default that
293 passes the string @option{-lgcc} to the linker.
295 @findex STARTFILE_SPEC
297 Another C string constant used much like @code{LINK_SPEC}. The
298 difference between the two is that @code{STARTFILE_SPEC} is used at
299 the very beginning of the command given to the linker.
301 If this macro is not defined, a default is provided that loads the
302 standard C startup file from the usual place. See @file{gcc.c}.
306 Another C string constant used much like @code{LINK_SPEC}. The
307 difference between the two is that @code{ENDFILE_SPEC} is used at
308 the very end of the command given to the linker.
310 Do not define this macro if it does not need to do anything.
312 @findex THREAD_MODEL_SPEC
313 @item THREAD_MODEL_SPEC
314 GCC @code{-v} will print the thread model GCC was configured to use.
315 However, this doesn't work on platforms that are multilibbed on thread
316 models, such as AIX 4.3. On such platforms, define
317 @code{THREAD_MODEL_SPEC} such that it evaluates to a string without
318 blanks that names one of the recognized thread models. @code{%*}, the
319 default value of this macro, will expand to the value of
320 @code{thread_file} set in @file{config.gcc}.
324 Define this macro to provide additional specifications to put in the
325 @file{specs} file that can be used in various specifications like
328 The definition should be an initializer for an array of structures,
329 containing a string constant, that defines the specification name, and a
330 string constant that provides the specification.
332 Do not define this macro if it does not need to do anything.
334 @code{EXTRA_SPECS} is useful when an architecture contains several
335 related targets, which have various @code{@dots{}_SPECS} which are similar
336 to each other, and the maintainer would like one central place to keep
339 For example, the PowerPC System V.4 targets use @code{EXTRA_SPECS} to
340 define either @code{_CALL_SYSV} when the System V calling sequence is
341 used or @code{_CALL_AIX} when the older AIX-based calling sequence is
344 The @file{config/rs6000/rs6000.h} target file defines:
347 #define EXTRA_SPECS \
348 @{ "cpp_sysv_default", CPP_SYSV_DEFAULT @},
350 #define CPP_SYS_DEFAULT ""
353 The @file{config/rs6000/sysv.h} target file defines:
357 "%@{posix: -D_POSIX_SOURCE @} \
358 %@{mcall-sysv: -D_CALL_SYSV @} %@{mcall-aix: -D_CALL_AIX @} \
359 %@{!mcall-sysv: %@{!mcall-aix: %(cpp_sysv_default) @}@} \
360 %@{msoft-float: -D_SOFT_FLOAT@} %@{mcpu=403: -D_SOFT_FLOAT@}"
362 #undef CPP_SYSV_DEFAULT
363 #define CPP_SYSV_DEFAULT "-D_CALL_SYSV"
366 while the @file{config/rs6000/eabiaix.h} target file defines
367 @code{CPP_SYSV_DEFAULT} as:
370 #undef CPP_SYSV_DEFAULT
371 #define CPP_SYSV_DEFAULT "-D_CALL_AIX"
374 @findex LINK_LIBGCC_SPECIAL
375 @item LINK_LIBGCC_SPECIAL
376 Define this macro if the driver program should find the library
377 @file{libgcc.a} itself and should not pass @option{-L} options to the
378 linker. If you do not define this macro, the driver program will pass
379 the argument @option{-lgcc} to tell the linker to do the search and will
380 pass @option{-L} options to it.
382 @findex LINK_LIBGCC_SPECIAL_1
383 @item LINK_LIBGCC_SPECIAL_1
384 Define this macro if the driver program should find the library
385 @file{libgcc.a}. If you do not define this macro, the driver program will pass
386 the argument @option{-lgcc} to tell the linker to do the search.
387 This macro is similar to @code{LINK_LIBGCC_SPECIAL}, except that it does
388 not affect @option{-L} options.
390 @findex LINK_COMMAND_SPEC
391 @item LINK_COMMAND_SPEC
392 A C string constant giving the complete command line need to execute the
393 linker. When you do this, you will need to update your port each time a
394 change is made to the link command line within @file{gcc.c}. Therefore,
395 define this macro only if you need to completely redefine the command
396 line for invoking the linker and there is no other way to accomplish
399 @findex LINK_ELIMINATE_DUPLICATE_LDIRECTORIES
400 @item LINK_ELIMINATE_DUPLICATE_LDIRECTORIES
401 A nonzero value causes @command{collect2} to remove duplicate @option{-L@var{directory}} search
402 directories from linking commands. Do not give it a nonzero value if
403 removing duplicate search directories changes the linker's semantics.
405 @findex MULTILIB_DEFAULTS
406 @item MULTILIB_DEFAULTS
407 Define this macro as a C expression for the initializer of an array of
408 string to tell the driver program which options are defaults for this
409 target and thus do not need to be handled specially when using
410 @code{MULTILIB_OPTIONS}.
412 Do not define this macro if @code{MULTILIB_OPTIONS} is not defined in
413 the target makefile fragment or if none of the options listed in
414 @code{MULTILIB_OPTIONS} are set by default.
415 @xref{Target Fragment}.
417 @findex RELATIVE_PREFIX_NOT_LINKDIR
418 @item RELATIVE_PREFIX_NOT_LINKDIR
419 Define this macro to tell @code{gcc} that it should only translate
420 a @option{-B} prefix into a @option{-L} linker option if the prefix
421 indicates an absolute file name.
423 @findex STANDARD_EXEC_PREFIX
424 @item STANDARD_EXEC_PREFIX
425 Define this macro as a C string constant if you wish to override the
426 standard choice of @file{/usr/local/lib/gcc-lib/} as the default prefix to
427 try when searching for the executable files of the compiler.
429 @findex MD_EXEC_PREFIX
431 If defined, this macro is an additional prefix to try after
432 @code{STANDARD_EXEC_PREFIX}. @code{MD_EXEC_PREFIX} is not searched
433 when the @option{-b} option is used, or the compiler is built as a cross
434 compiler. If you define @code{MD_EXEC_PREFIX}, then be sure to add it
435 to the list of directories used to find the assembler in @file{configure.in}.
437 @findex STANDARD_STARTFILE_PREFIX
438 @item STANDARD_STARTFILE_PREFIX
439 Define this macro as a C string constant if you wish to override the
440 standard choice of @file{/usr/local/lib/} as the default prefix to
441 try when searching for startup files such as @file{crt0.o}.
443 @findex MD_STARTFILE_PREFIX
444 @item MD_STARTFILE_PREFIX
445 If defined, this macro supplies an additional prefix to try after the
446 standard prefixes. @code{MD_EXEC_PREFIX} is not searched when the
447 @option{-b} option is used, or when the compiler is built as a cross
450 @findex MD_STARTFILE_PREFIX_1
451 @item MD_STARTFILE_PREFIX_1
452 If defined, this macro supplies yet another prefix to try after the
453 standard prefixes. It is not searched when the @option{-b} option is
454 used, or when the compiler is built as a cross compiler.
456 @findex INIT_ENVIRONMENT
457 @item INIT_ENVIRONMENT
458 Define this macro as a C string constant if you wish to set environment
459 variables for programs called by the driver, such as the assembler and
460 loader. The driver passes the value of this macro to @code{putenv} to
461 initialize the necessary environment variables.
463 @findex LOCAL_INCLUDE_DIR
464 @item LOCAL_INCLUDE_DIR
465 Define this macro as a C string constant if you wish to override the
466 standard choice of @file{/usr/local/include} as the default prefix to
467 try when searching for local header files. @code{LOCAL_INCLUDE_DIR}
468 comes before @code{SYSTEM_INCLUDE_DIR} in the search order.
470 Cross compilers do not search either @file{/usr/local/include} or its
473 @findex MODIFY_TARGET_NAME
474 @item MODIFY_TARGET_NAME
475 Define this macro if you with to define command-line switches that modify the
478 For each switch, you can include a string to be appended to the first
479 part of the configuration name or a string to be deleted from the
480 configuration name, if present. The definition should be an initializer
481 for an array of structures. Each array element should have three
482 elements: the switch name (a string constant, including the initial
483 dash), one of the enumeration codes @code{ADD} or @code{DELETE} to
484 indicate whether the string should be inserted or deleted, and the string
485 to be inserted or deleted (a string constant).
487 For example, on a machine where @samp{64} at the end of the
488 configuration name denotes a 64-bit target and you want the @option{-32}
489 and @option{-64} switches to select between 32- and 64-bit targets, you would
493 #define MODIFY_TARGET_NAME \
494 @{ @{ "-32", DELETE, "64"@}, \
495 @{"-64", ADD, "64"@}@}
499 @findex SYSTEM_INCLUDE_DIR
500 @item SYSTEM_INCLUDE_DIR
501 Define this macro as a C string constant if you wish to specify a
502 system-specific directory to search for header files before the standard
503 directory. @code{SYSTEM_INCLUDE_DIR} comes before
504 @code{STANDARD_INCLUDE_DIR} in the search order.
506 Cross compilers do not use this macro and do not search the directory
509 @findex STANDARD_INCLUDE_DIR
510 @item STANDARD_INCLUDE_DIR
511 Define this macro as a C string constant if you wish to override the
512 standard choice of @file{/usr/include} as the default prefix to
513 try when searching for header files.
515 Cross compilers do not use this macro and do not search either
516 @file{/usr/include} or its replacement.
518 @findex STANDARD_INCLUDE_COMPONENT
519 @item STANDARD_INCLUDE_COMPONENT
520 The ``component'' corresponding to @code{STANDARD_INCLUDE_DIR}.
521 See @code{INCLUDE_DEFAULTS}, below, for the description of components.
522 If you do not define this macro, no component is used.
524 @findex INCLUDE_DEFAULTS
525 @item INCLUDE_DEFAULTS
526 Define this macro if you wish to override the entire default search path
527 for include files. For a native compiler, the default search path
528 usually consists of @code{GCC_INCLUDE_DIR}, @code{LOCAL_INCLUDE_DIR},
529 @code{SYSTEM_INCLUDE_DIR}, @code{GPLUSPLUS_INCLUDE_DIR}, and
530 @code{STANDARD_INCLUDE_DIR}. In addition, @code{GPLUSPLUS_INCLUDE_DIR}
531 and @code{GCC_INCLUDE_DIR} are defined automatically by @file{Makefile},
532 and specify private search areas for GCC@. The directory
533 @code{GPLUSPLUS_INCLUDE_DIR} is used only for C++ programs.
535 The definition should be an initializer for an array of structures.
536 Each array element should have four elements: the directory name (a
537 string constant), the component name (also a string constant), a flag
538 for C++-only directories,
539 and a flag showing that the includes in the directory don't need to be
540 wrapped in @code{extern @samp{C}} when compiling C++. Mark the end of
541 the array with a null element.
543 The component name denotes what GNU package the include file is part of,
544 if any, in all upper-case letters. For example, it might be @samp{GCC}
545 or @samp{BINUTILS}. If the package is part of a vendor-supplied
546 operating system, code the component name as @samp{0}.
548 For example, here is the definition used for VAX/VMS:
551 #define INCLUDE_DEFAULTS \
553 @{ "GNU_GXX_INCLUDE:", "G++", 1, 1@}, \
554 @{ "GNU_CC_INCLUDE:", "GCC", 0, 0@}, \
555 @{ "SYS$SYSROOT:[SYSLIB.]", 0, 0, 0@}, \
562 Here is the order of prefixes tried for exec files:
566 Any prefixes specified by the user with @option{-B}.
569 The environment variable @code{GCC_EXEC_PREFIX}, if any.
572 The directories specified by the environment variable @code{COMPILER_PATH}.
575 The macro @code{STANDARD_EXEC_PREFIX}.
578 @file{/usr/lib/gcc/}.
581 The macro @code{MD_EXEC_PREFIX}, if any.
584 Here is the order of prefixes tried for startfiles:
588 Any prefixes specified by the user with @option{-B}.
591 The environment variable @code{GCC_EXEC_PREFIX}, if any.
594 The directories specified by the environment variable @code{LIBRARY_PATH}
595 (or port-specific name; native only, cross compilers do not use this).
598 The macro @code{STANDARD_EXEC_PREFIX}.
601 @file{/usr/lib/gcc/}.
604 The macro @code{MD_EXEC_PREFIX}, if any.
607 The macro @code{MD_STARTFILE_PREFIX}, if any.
610 The macro @code{STANDARD_STARTFILE_PREFIX}.
619 @node Run-time Target
620 @section Run-time Target Specification
621 @cindex run-time target specification
622 @cindex predefined macros
623 @cindex target specifications
625 @c prevent bad page break with this line
626 Here are run-time target specifications.
629 @findex CPP_PREDEFINES
631 Define this to be a string constant containing @option{-D} options to
632 define the predefined macros that identify this machine and system.
633 These macros will be predefined unless the @option{-ansi} option (or a
634 @option{-std} option for strict ISO C conformance) is specified.
636 In addition, a parallel set of macros are predefined, whose names are
637 made by appending @samp{__} at the beginning and at the end. These
638 @samp{__} macros are permitted by the ISO standard, so they are
639 predefined regardless of whether @option{-ansi} or a @option{-std} option
642 For example, on the Sun, one can use the following value:
645 "-Dmc68000 -Dsun -Dunix"
648 The result is to define the macros @code{__mc68000__}, @code{__sun__}
649 and @code{__unix__} unconditionally, and the macros @code{mc68000},
650 @code{sun} and @code{unix} provided @option{-ansi} is not specified.
652 @findex extern int target_flags
653 @item extern int target_flags;
654 This declaration should be present.
656 @cindex optional hardware or system features
657 @cindex features, optional, in system conventions
659 This series of macros is to allow compiler command arguments to
660 enable or disable the use of optional features of the target machine.
661 For example, one machine description serves both the 68000 and
662 the 68020; a command argument tells the compiler whether it should
663 use 68020-only instructions or not. This command argument works
664 by means of a macro @code{TARGET_68020} that tests a bit in
667 Define a macro @code{TARGET_@var{featurename}} for each such option.
668 Its definition should test a bit in @code{target_flags}. It is
669 recommended that a helper macro @code{TARGET_MASK_@var{featurename}}
670 is defined for each bit-value to test, and used in
671 @code{TARGET_@var{featurename}} and @code{TARGET_SWITCHES}. For
675 #define TARGET_MASK_68020 1
676 #define TARGET_68020 (target_flags & TARGET_MASK_68020)
679 One place where these macros are used is in the condition-expressions
680 of instruction patterns. Note how @code{TARGET_68020} appears
681 frequently in the 68000 machine description file, @file{m68k.md}.
682 Another place they are used is in the definitions of the other
683 macros in the @file{@var{machine}.h} file.
685 @findex TARGET_SWITCHES
686 @item TARGET_SWITCHES
687 This macro defines names of command options to set and clear
688 bits in @code{target_flags}. Its definition is an initializer
689 with a subgrouping for each command option.
691 Each subgrouping contains a string constant, that defines the option
692 name, a number, which contains the bits to set in
693 @code{target_flags}, and a second string which is the description
694 displayed by @option{--help}. If the number is negative then the bits specified
695 by the number are cleared instead of being set. If the description
696 string is present but empty, then no help information will be displayed
697 for that option, but it will not count as an undocumented option. The
698 actual option name is made by appending @samp{-m} to the specified name.
699 Non-empty description strings should be marked with @code{N_(@dots{})} for
700 @command{xgettext}. In addition to the description for @option{--help},
701 more detailed documentation for each option should be added to
704 One of the subgroupings should have a null string. The number in
705 this grouping is the default value for @code{target_flags}. Any
706 target options act starting with that value.
708 Here is an example which defines @option{-m68000} and @option{-m68020}
709 with opposite meanings, and picks the latter as the default:
712 #define TARGET_SWITCHES \
713 @{ @{ "68020", TARGET_MASK_68020, "" @}, \
714 @{ "68000", -TARGET_MASK_68020, \
715 N_("Compile for the 68000") @}, \
716 @{ "", TARGET_MASK_68020, "" @}@}
719 @findex TARGET_OPTIONS
721 This macro is similar to @code{TARGET_SWITCHES} but defines names of command
722 options that have values. Its definition is an initializer with a
723 subgrouping for each command option.
725 Each subgrouping contains a string constant, that defines the fixed part
726 of the option name, the address of a variable, and a description string
727 (which should again be marked with @code{N_(@dots{})}).
728 The variable, type @code{char *}, is set to the variable part of the
729 given option if the fixed part matches. The actual option name is made
730 by appending @samp{-m} to the specified name. Again, each option should
731 also be documented in @file{invoke.texi}.
733 Here is an example which defines @option{-mshort-data-@var{number}}. If the
734 given option is @option{-mshort-data-512}, the variable @code{m88k_short_data}
735 will be set to the string @code{"512"}.
738 extern char *m88k_short_data;
739 #define TARGET_OPTIONS \
740 @{ @{ "short-data-", &m88k_short_data, \
741 N_("Specify the size of the short data section") @} @}
744 @findex TARGET_VERSION
746 This macro is a C statement to print on @code{stderr} a string
747 describing the particular machine description choice. Every machine
748 description should define @code{TARGET_VERSION}. For example:
752 #define TARGET_VERSION \
753 fprintf (stderr, " (68k, Motorola syntax)");
755 #define TARGET_VERSION \
756 fprintf (stderr, " (68k, MIT syntax)");
760 @findex OVERRIDE_OPTIONS
761 @item OVERRIDE_OPTIONS
762 Sometimes certain combinations of command options do not make sense on
763 a particular target machine. You can define a macro
764 @code{OVERRIDE_OPTIONS} to take account of this. This macro, if
765 defined, is executed once just after all the command options have been
768 Don't use this macro to turn on various extra optimizations for
769 @option{-O}. That is what @code{OPTIMIZATION_OPTIONS} is for.
771 @findex OPTIMIZATION_OPTIONS
772 @item OPTIMIZATION_OPTIONS (@var{level}, @var{size})
773 Some machines may desire to change what optimizations are performed for
774 various optimization levels. This macro, if defined, is executed once
775 just after the optimization level is determined and before the remainder
776 of the command options have been parsed. Values set in this macro are
777 used as the default values for the other command line options.
779 @var{level} is the optimization level specified; 2 if @option{-O2} is
780 specified, 1 if @option{-O} is specified, and 0 if neither is specified.
782 @var{size} is nonzero if @option{-Os} is specified and zero otherwise.
784 You should not use this macro to change options that are not
785 machine-specific. These should uniformly selected by the same
786 optimization level on all supported machines. Use this macro to enable
787 machine-specific optimizations.
789 @strong{Do not examine @code{write_symbols} in
790 this macro!} The debugging options are not supposed to alter the
793 @findex CAN_DEBUG_WITHOUT_FP
794 @item CAN_DEBUG_WITHOUT_FP
795 Define this macro if debugging can be performed even without a frame
796 pointer. If this macro is defined, GCC will turn on the
797 @option{-fomit-frame-pointer} option whenever @option{-O} is specified.
800 @node Per-Function Data
801 @section Defining data structures for per-function information.
802 @cindex per-function data
803 @cindex data structures
805 If the target needs to store information on a per-function basis, GCC
806 provides a macro and a couple of variables to allow this. Note, just
807 using statics to store the information is a bad idea, since GCC supports
808 nested functions, so you can be halfway through encoding one function
809 when another one comes along.
811 GCC defines a data structure called @code{struct function} which
812 contains all of the data specific to an individual function. This
813 structure contains a field called @code{machine} whose type is
814 @code{struct machine_function *}, which can be used by targets to point
815 to their own specific data.
817 If a target needs per-function specific data it should define the type
818 @code{struct machine_function} and also the macro
819 @code{INIT_EXPANDERS}. This macro should be used to initialize some or
820 all of the function pointers @code{init_machine_status},
821 @code{free_machine_status} and @code{mark_machine_status}. These
822 pointers are explained below.
824 One typical use of per-function, target specific data is to create an
825 RTX to hold the register containing the function's return address. This
826 RTX can then be used to implement the @code{__builtin_return_address}
827 function, for level 0.
829 Note---earlier implementations of GCC used a single data area to hold
830 all of the per-function information. Thus when processing of a nested
831 function began the old per-function data had to be pushed onto a
832 stack, and when the processing was finished, it had to be popped off the
833 stack. GCC used to provide function pointers called
834 @code{save_machine_status} and @code{restore_machine_status} to handle
835 the saving and restoring of the target specific information. Since the
836 single data area approach is no longer used, these pointers are no
839 The macro and function pointers are described below.
842 @findex INIT_EXPANDERS
844 Macro called to initialize any target specific information. This macro
845 is called once per function, before generation of any RTL has begun.
846 The intention of this macro is to allow the initialization of the
847 function pointers below.
849 @findex init_machine_status
850 @item init_machine_status
851 This is a @code{void (*)(struct function *)} function pointer. If this
852 pointer is non-@code{NULL} it will be called once per function, before function
853 compilation starts, in order to allow the target to perform any target
854 specific initialization of the @code{struct function} structure. It is
855 intended that this would be used to initialize the @code{machine} of
858 @findex free_machine_status
859 @item free_machine_status
860 This is a @code{void (*)(struct function *)} function pointer. If this
861 pointer is non-@code{NULL} it will be called once per function, after the
862 function has been compiled, in order to allow any memory allocated
863 during the @code{init_machine_status} function call to be freed.
865 @findex mark_machine_status
866 @item mark_machine_status
867 This is a @code{void (*)(struct function *)} function pointer. If this
868 pointer is non-@code{NULL} it will be called once per function in order to mark
869 any data items in the @code{struct machine_function} structure which
870 need garbage collection.
875 @section Storage Layout
876 @cindex storage layout
878 Note that the definitions of the macros in this table which are sizes or
879 alignments measured in bits do not need to be constant. They can be C
880 expressions that refer to static variables, such as the @code{target_flags}.
881 @xref{Run-time Target}.
884 @findex BITS_BIG_ENDIAN
885 @item BITS_BIG_ENDIAN
886 Define this macro to have the value 1 if the most significant bit in a
887 byte has the lowest number; otherwise define it to have the value zero.
888 This means that bit-field instructions count from the most significant
889 bit. If the machine has no bit-field instructions, then this must still
890 be defined, but it doesn't matter which value it is defined to. This
891 macro need not be a constant.
893 This macro does not affect the way structure fields are packed into
894 bytes or words; that is controlled by @code{BYTES_BIG_ENDIAN}.
896 @findex BYTES_BIG_ENDIAN
897 @item BYTES_BIG_ENDIAN
898 Define this macro to have the value 1 if the most significant byte in a
899 word has the lowest number. This macro need not be a constant.
901 @findex WORDS_BIG_ENDIAN
902 @item WORDS_BIG_ENDIAN
903 Define this macro to have the value 1 if, in a multiword object, the
904 most significant word has the lowest number. This applies to both
905 memory locations and registers; GCC fundamentally assumes that the
906 order of words in memory is the same as the order in registers. This
907 macro need not be a constant.
909 @findex LIBGCC2_WORDS_BIG_ENDIAN
910 @item LIBGCC2_WORDS_BIG_ENDIAN
911 Define this macro if @code{WORDS_BIG_ENDIAN} is not constant. This must be a
912 constant value with the same meaning as @code{WORDS_BIG_ENDIAN}, which will be
913 used only when compiling @file{libgcc2.c}. Typically the value will be set
914 based on preprocessor defines.
916 @findex FLOAT_WORDS_BIG_ENDIAN
917 @item FLOAT_WORDS_BIG_ENDIAN
918 Define this macro to have the value 1 if @code{DFmode}, @code{XFmode} or
919 @code{TFmode} floating point numbers are stored in memory with the word
920 containing the sign bit at the lowest address; otherwise define it to
921 have the value 0. This macro need not be a constant.
923 You need not define this macro if the ordering is the same as for
926 @findex BITS_PER_UNIT
928 Define this macro to be the number of bits in an addressable storage
929 unit (byte); normally 8.
931 @findex BITS_PER_WORD
933 Number of bits in a word; normally 32.
935 @findex MAX_BITS_PER_WORD
936 @item MAX_BITS_PER_WORD
937 Maximum number of bits in a word. If this is undefined, the default is
938 @code{BITS_PER_WORD}. Otherwise, it is the constant value that is the
939 largest value that @code{BITS_PER_WORD} can have at run-time.
941 @findex UNITS_PER_WORD
943 Number of storage units in a word; normally 4.
945 @findex MIN_UNITS_PER_WORD
946 @item MIN_UNITS_PER_WORD
947 Minimum number of units in a word. If this is undefined, the default is
948 @code{UNITS_PER_WORD}. Otherwise, it is the constant value that is the
949 smallest value that @code{UNITS_PER_WORD} can have at run-time.
953 Width of a pointer, in bits. You must specify a value no wider than the
954 width of @code{Pmode}. If it is not equal to the width of @code{Pmode},
955 you must define @code{POINTERS_EXTEND_UNSIGNED}.
957 @findex POINTERS_EXTEND_UNSIGNED
958 @item POINTERS_EXTEND_UNSIGNED
959 A C expression whose value is greater than zero if pointers that need to be
960 extended from being @code{POINTER_SIZE} bits wide to @code{Pmode} are to
961 be zero-extended and zero if they are to be sign-extended. If the value
962 is less then zero then there must be an "ptr_extend" instruction that
963 extends a pointer from @code{POINTER_SIZE} to @code{Pmode}.
965 You need not define this macro if the @code{POINTER_SIZE} is equal
966 to the width of @code{Pmode}.
969 @item PROMOTE_MODE (@var{m}, @var{unsignedp}, @var{type})
970 A macro to update @var{m} and @var{unsignedp} when an object whose type
971 is @var{type} and which has the specified mode and signedness is to be
972 stored in a register. This macro is only called when @var{type} is a
975 On most RISC machines, which only have operations that operate on a full
976 register, define this macro to set @var{m} to @code{word_mode} if
977 @var{m} is an integer mode narrower than @code{BITS_PER_WORD}. In most
978 cases, only integer modes should be widened because wider-precision
979 floating-point operations are usually more expensive than their narrower
982 For most machines, the macro definition does not change @var{unsignedp}.
983 However, some machines, have instructions that preferentially handle
984 either signed or unsigned quantities of certain modes. For example, on
985 the DEC Alpha, 32-bit loads from memory and 32-bit add instructions
986 sign-extend the result to 64 bits. On such machines, set
987 @var{unsignedp} according to which kind of extension is more efficient.
989 Do not define this macro if it would never modify @var{m}.
991 @findex PROMOTE_FUNCTION_ARGS
992 @item PROMOTE_FUNCTION_ARGS
993 Define this macro if the promotion described by @code{PROMOTE_MODE}
994 should also be done for outgoing function arguments.
996 @findex PROMOTE_FUNCTION_RETURN
997 @item PROMOTE_FUNCTION_RETURN
998 Define this macro if the promotion described by @code{PROMOTE_MODE}
999 should also be done for the return value of functions.
1001 If this macro is defined, @code{FUNCTION_VALUE} must perform the same
1002 promotions done by @code{PROMOTE_MODE}.
1004 @findex PROMOTE_FOR_CALL_ONLY
1005 @item PROMOTE_FOR_CALL_ONLY
1006 Define this macro if the promotion described by @code{PROMOTE_MODE}
1007 should @emph{only} be performed for outgoing function arguments or
1008 function return values, as specified by @code{PROMOTE_FUNCTION_ARGS}
1009 and @code{PROMOTE_FUNCTION_RETURN}, respectively.
1011 @findex PARM_BOUNDARY
1013 Normal alignment required for function parameters on the stack, in
1014 bits. All stack parameters receive at least this much alignment
1015 regardless of data type. On most machines, this is the same as the
1018 @findex STACK_BOUNDARY
1019 @item STACK_BOUNDARY
1020 Define this macro to the minimum alignment enforced by hardware for the
1021 stack pointer on this machine. The definition is a C expression for the
1022 desired alignment (measured in bits). This value is used as a default
1023 if @code{PREFERRED_STACK_BOUNDARY} is not defined. On most machines,
1024 this should be the same as @code{PARM_BOUNDARY}.
1026 @findex PREFERRED_STACK_BOUNDARY
1027 @item PREFERRED_STACK_BOUNDARY
1028 Define this macro if you wish to preserve a certain alignment for the
1029 stack pointer, greater than what the hardware enforces. The definition
1030 is a C expression for the desired alignment (measured in bits). This
1031 macro must evaluate to a value equal to or larger than
1032 @code{STACK_BOUNDARY}.
1034 @findex FORCE_PREFERRED_STACK_BOUNDARY_IN_MAIN
1035 @item FORCE_PREFERRED_STACK_BOUNDARY_IN_MAIN
1036 A C expression that evaluates true if @code{PREFERRED_STACK_BOUNDARY} is
1037 not guaranteed by the runtime and we should emit code to align the stack
1038 at the beginning of @code{main}.
1040 @cindex @code{PUSH_ROUNDING}, interaction with @code{PREFERRED_STACK_BOUNDARY}
1041 If @code{PUSH_ROUNDING} is not defined, the stack will always be aligned
1042 to the specified boundary. If @code{PUSH_ROUNDING} is defined and specifies
1043 a less strict alignment than @code{PREFERRED_STACK_BOUNDARY}, the stack may
1044 be momentarily unaligned while pushing arguments.
1046 @findex FUNCTION_BOUNDARY
1047 @item FUNCTION_BOUNDARY
1048 Alignment required for a function entry point, in bits.
1050 @findex BIGGEST_ALIGNMENT
1051 @item BIGGEST_ALIGNMENT
1052 Biggest alignment that any data type can require on this machine, in bits.
1054 @findex MINIMUM_ATOMIC_ALIGNMENT
1055 @item MINIMUM_ATOMIC_ALIGNMENT
1056 If defined, the smallest alignment, in bits, that can be given to an
1057 object that can be referenced in one operation, without disturbing any
1058 nearby object. Normally, this is @code{BITS_PER_UNIT}, but may be larger
1059 on machines that don't have byte or half-word store operations.
1061 @findex BIGGEST_FIELD_ALIGNMENT
1062 @item BIGGEST_FIELD_ALIGNMENT
1063 Biggest alignment that any structure or union field can require on this
1064 machine, in bits. If defined, this overrides @code{BIGGEST_ALIGNMENT} for
1065 structure and union fields only, unless the field alignment has been set
1066 by the @code{__attribute__ ((aligned (@var{n})))} construct.
1068 @findex ADJUST_FIELD_ALIGN
1069 @item ADJUST_FIELD_ALIGN (@var{field}, @var{computed})
1070 An expression for the alignment of a structure field @var{field} if the
1071 alignment computed in the usual way is @var{computed}. GCC uses
1072 this value instead of the value in @code{BIGGEST_ALIGNMENT} or
1073 @code{BIGGEST_FIELD_ALIGNMENT}, if defined.
1075 @findex MAX_OFILE_ALIGNMENT
1076 @item MAX_OFILE_ALIGNMENT
1077 Biggest alignment supported by the object file format of this machine.
1078 Use this macro to limit the alignment which can be specified using the
1079 @code{__attribute__ ((aligned (@var{n})))} construct. If not defined,
1080 the default value is @code{BIGGEST_ALIGNMENT}.
1082 @findex DATA_ALIGNMENT
1083 @item DATA_ALIGNMENT (@var{type}, @var{basic-align})
1084 If defined, a C expression to compute the alignment for a variable in
1085 the static store. @var{type} is the data type, and @var{basic-align} is
1086 the alignment that the object would ordinarily have. The value of this
1087 macro is used instead of that alignment to align the object.
1089 If this macro is not defined, then @var{basic-align} is used.
1092 One use of this macro is to increase alignment of medium-size data to
1093 make it all fit in fewer cache lines. Another is to cause character
1094 arrays to be word-aligned so that @code{strcpy} calls that copy
1095 constants to character arrays can be done inline.
1097 @findex CONSTANT_ALIGNMENT
1098 @item CONSTANT_ALIGNMENT (@var{constant}, @var{basic-align})
1099 If defined, a C expression to compute the alignment given to a constant
1100 that is being placed in memory. @var{constant} is the constant and
1101 @var{basic-align} is the alignment that the object would ordinarily
1102 have. The value of this macro is used instead of that alignment to
1105 If this macro is not defined, then @var{basic-align} is used.
1107 The typical use of this macro is to increase alignment for string
1108 constants to be word aligned so that @code{strcpy} calls that copy
1109 constants can be done inline.
1111 @findex LOCAL_ALIGNMENT
1112 @item LOCAL_ALIGNMENT (@var{type}, @var{basic-align})
1113 If defined, a C expression to compute the alignment for a variable in
1114 the local store. @var{type} is the data type, and @var{basic-align} is
1115 the alignment that the object would ordinarily have. The value of this
1116 macro is used instead of that alignment to align the object.
1118 If this macro is not defined, then @var{basic-align} is used.
1120 One use of this macro is to increase alignment of medium-size data to
1121 make it all fit in fewer cache lines.
1123 @findex EMPTY_FIELD_BOUNDARY
1124 @item EMPTY_FIELD_BOUNDARY
1125 Alignment in bits to be given to a structure bit-field that follows an
1126 empty field such as @code{int : 0;}.
1128 Note that @code{PCC_BITFIELD_TYPE_MATTERS} also affects the alignment
1129 that results from an empty field.
1131 @findex STRUCTURE_SIZE_BOUNDARY
1132 @item STRUCTURE_SIZE_BOUNDARY
1133 Number of bits which any structure or union's size must be a multiple of.
1134 Each structure or union's size is rounded up to a multiple of this.
1136 If you do not define this macro, the default is the same as
1137 @code{BITS_PER_UNIT}.
1139 @findex STRICT_ALIGNMENT
1140 @item STRICT_ALIGNMENT
1141 Define this macro to be the value 1 if instructions will fail to work
1142 if given data not on the nominal alignment. If instructions will merely
1143 go slower in that case, define this macro as 0.
1145 @findex PCC_BITFIELD_TYPE_MATTERS
1146 @item PCC_BITFIELD_TYPE_MATTERS
1147 Define this if you wish to imitate the way many other C compilers handle
1148 alignment of bit-fields and the structures that contain them.
1150 The behavior is that the type written for a bit-field (@code{int},
1151 @code{short}, or other integer type) imposes an alignment for the
1152 entire structure, as if the structure really did contain an ordinary
1153 field of that type. In addition, the bit-field is placed within the
1154 structure so that it would fit within such a field, not crossing a
1157 Thus, on most machines, a bit-field whose type is written as @code{int}
1158 would not cross a four-byte boundary, and would force four-byte
1159 alignment for the whole structure. (The alignment used may not be four
1160 bytes; it is controlled by the other alignment parameters.)
1162 If the macro is defined, its definition should be a C expression;
1163 a nonzero value for the expression enables this behavior.
1165 Note that if this macro is not defined, or its value is zero, some
1166 bit-fields may cross more than one alignment boundary. The compiler can
1167 support such references if there are @samp{insv}, @samp{extv}, and
1168 @samp{extzv} insns that can directly reference memory.
1170 The other known way of making bit-fields work is to define
1171 @code{STRUCTURE_SIZE_BOUNDARY} as large as @code{BIGGEST_ALIGNMENT}.
1172 Then every structure can be accessed with fullwords.
1174 Unless the machine has bit-field instructions or you define
1175 @code{STRUCTURE_SIZE_BOUNDARY} that way, you must define
1176 @code{PCC_BITFIELD_TYPE_MATTERS} to have a nonzero value.
1178 If your aim is to make GCC use the same conventions for laying out
1179 bit-fields as are used by another compiler, here is how to investigate
1180 what the other compiler does. Compile and run this program:
1199 printf ("Size of foo1 is %d\n",
1200 sizeof (struct foo1));
1201 printf ("Size of foo2 is %d\n",
1202 sizeof (struct foo2));
1207 If this prints 2 and 5, then the compiler's behavior is what you would
1208 get from @code{PCC_BITFIELD_TYPE_MATTERS}.
1210 @findex BITFIELD_NBYTES_LIMITED
1211 @item BITFIELD_NBYTES_LIMITED
1212 Like PCC_BITFIELD_TYPE_MATTERS except that its effect is limited to
1213 aligning a bit-field within the structure.
1215 @findex MEMBER_TYPE_FORCES_BLK
1216 @item MEMBER_TYPE_FORCES_BLK (@var{field})
1217 Return 1 if a structure or array containing @var{field} should be accessed using
1220 Normally, this is not needed. See the file @file{c4x.h} for an example
1221 of how to use this macro to prevent a structure having a floating point
1222 field from being accessed in an integer mode.
1224 @findex ROUND_TYPE_SIZE
1225 @item ROUND_TYPE_SIZE (@var{type}, @var{computed}, @var{specified})
1226 Define this macro as an expression for the overall size of a type
1227 (given by @var{type} as a tree node) when the size computed in the
1228 usual way is @var{computed} and the alignment is @var{specified}.
1230 The default is to round @var{computed} up to a multiple of @var{specified}.
1232 @findex ROUND_TYPE_SIZE_UNIT
1233 @item ROUND_TYPE_SIZE_UNIT (@var{type}, @var{computed}, @var{specified})
1234 Similar to @code{ROUND_TYPE_SIZE}, but sizes and alignments are
1235 specified in units (bytes). If you define @code{ROUND_TYPE_SIZE},
1236 you must also define this macro and they must be defined consistently
1239 @findex ROUND_TYPE_ALIGN
1240 @item ROUND_TYPE_ALIGN (@var{type}, @var{computed}, @var{specified})
1241 Define this macro as an expression for the alignment of a type (given
1242 by @var{type} as a tree node) if the alignment computed in the usual
1243 way is @var{computed} and the alignment explicitly specified was
1246 The default is to use @var{specified} if it is larger; otherwise, use
1247 the smaller of @var{computed} and @code{BIGGEST_ALIGNMENT}
1249 @findex MAX_FIXED_MODE_SIZE
1250 @item MAX_FIXED_MODE_SIZE
1251 An integer expression for the size in bits of the largest integer
1252 machine mode that should actually be used. All integer machine modes of
1253 this size or smaller can be used for structures and unions with the
1254 appropriate sizes. If this macro is undefined, @code{GET_MODE_BITSIZE
1255 (DImode)} is assumed.
1257 @findex VECTOR_MODE_SUPPORTED_P
1258 @item VECTOR_MODE_SUPPORTED_P(@var{mode})
1259 Define this macro to be nonzero if the port is prepared to handle insns
1260 involving vector mode @var{mode}. At the very least, it must have move
1261 patterns for this mode.
1263 @findex STACK_SAVEAREA_MODE
1264 @item STACK_SAVEAREA_MODE (@var{save_level})
1265 If defined, an expression of type @code{enum machine_mode} that
1266 specifies the mode of the save area operand of a
1267 @code{save_stack_@var{level}} named pattern (@pxref{Standard Names}).
1268 @var{save_level} is one of @code{SAVE_BLOCK}, @code{SAVE_FUNCTION}, or
1269 @code{SAVE_NONLOCAL} and selects which of the three named patterns is
1270 having its mode specified.
1272 You need not define this macro if it always returns @code{Pmode}. You
1273 would most commonly define this macro if the
1274 @code{save_stack_@var{level}} patterns need to support both a 32- and a
1277 @findex STACK_SIZE_MODE
1278 @item STACK_SIZE_MODE
1279 If defined, an expression of type @code{enum machine_mode} that
1280 specifies the mode of the size increment operand of an
1281 @code{allocate_stack} named pattern (@pxref{Standard Names}).
1283 You need not define this macro if it always returns @code{word_mode}.
1284 You would most commonly define this macro if the @code{allocate_stack}
1285 pattern needs to support both a 32- and a 64-bit mode.
1287 @findex CHECK_FLOAT_VALUE
1288 @item CHECK_FLOAT_VALUE (@var{mode}, @var{value}, @var{overflow})
1289 A C statement to validate the value @var{value} (of type
1290 @code{double}) for mode @var{mode}. This means that you check whether
1291 @var{value} fits within the possible range of values for mode
1292 @var{mode} on this target machine. The mode @var{mode} is always
1293 a mode of class @code{MODE_FLOAT}. @var{overflow} is nonzero if
1294 the value is already known to be out of range.
1296 If @var{value} is not valid or if @var{overflow} is nonzero, you should
1297 set @var{overflow} to 1 and then assign some valid value to @var{value}.
1298 Allowing an invalid value to go through the compiler can produce
1299 incorrect assembler code which may even cause Unix assemblers to crash.
1301 This macro need not be defined if there is no work for it to do.
1303 @findex TARGET_FLOAT_FORMAT
1304 @item TARGET_FLOAT_FORMAT
1305 A code distinguishing the floating point format of the target machine.
1306 There are five defined values:
1309 @findex IEEE_FLOAT_FORMAT
1310 @item IEEE_FLOAT_FORMAT
1311 This code indicates IEEE floating point. It is the default; there is no
1312 need to define this macro when the format is IEEE@.
1314 @findex VAX_FLOAT_FORMAT
1315 @item VAX_FLOAT_FORMAT
1316 This code indicates the peculiar format used on the VAX.
1318 @findex IBM_FLOAT_FORMAT
1319 @item IBM_FLOAT_FORMAT
1320 This code indicates the format used on the IBM System/370.
1322 @findex C4X_FLOAT_FORMAT
1323 @item C4X_FLOAT_FORMAT
1324 This code indicates the format used on the TMS320C3x/C4x.
1326 @findex UNKNOWN_FLOAT_FORMAT
1327 @item UNKNOWN_FLOAT_FORMAT
1328 This code indicates any other format.
1331 The value of this macro is compared with @code{HOST_FLOAT_FORMAT}
1332 (@pxref{Config}) to determine whether the target machine has the same
1333 format as the host machine. If any other formats are actually in use on
1334 supported machines, new codes should be defined for them.
1336 The ordering of the component words of floating point values stored in
1337 memory is controlled by @code{FLOAT_WORDS_BIG_ENDIAN} for the target
1338 machine and @code{HOST_FLOAT_WORDS_BIG_ENDIAN} for the host.
1343 @section Layout of Source Language Data Types
1345 These macros define the sizes and other characteristics of the standard
1346 basic data types used in programs being compiled. Unlike the macros in
1347 the previous section, these apply to specific features of C and related
1348 languages, rather than to fundamental aspects of storage layout.
1351 @findex INT_TYPE_SIZE
1353 A C expression for the size in bits of the type @code{int} on the
1354 target machine. If you don't define this, the default is one word.
1356 @findex SHORT_TYPE_SIZE
1357 @item SHORT_TYPE_SIZE
1358 A C expression for the size in bits of the type @code{short} on the
1359 target machine. If you don't define this, the default is half a word.
1360 (If this would be less than one storage unit, it is rounded up to one
1363 @findex LONG_TYPE_SIZE
1364 @item LONG_TYPE_SIZE
1365 A C expression for the size in bits of the type @code{long} on the
1366 target machine. If you don't define this, the default is one word.
1368 @findex ADA_LONG_TYPE_SIZE
1369 @item ADA_LONG_TYPE_SIZE
1370 On some machines, the size used for the Ada equivalent of the type
1371 @code{long} by a native Ada compiler differs from that used by C. In
1372 that situation, define this macro to be a C expression to be used for
1373 the size of that type. If you don't define this, the default is the
1374 value of @code{LONG_TYPE_SIZE}.
1376 @findex MAX_LONG_TYPE_SIZE
1377 @item MAX_LONG_TYPE_SIZE
1378 Maximum number for the size in bits of the type @code{long} on the
1379 target machine. If this is undefined, the default is
1380 @code{LONG_TYPE_SIZE}. Otherwise, it is the constant value that is the
1381 largest value that @code{LONG_TYPE_SIZE} can have at run-time. This is
1384 @findex LONG_LONG_TYPE_SIZE
1385 @item LONG_LONG_TYPE_SIZE
1386 A C expression for the size in bits of the type @code{long long} on the
1387 target machine. If you don't define this, the default is two
1388 words. If you want to support GNU Ada on your machine, the value of this
1389 macro must be at least 64.
1391 @findex CHAR_TYPE_SIZE
1392 @item CHAR_TYPE_SIZE
1393 A C expression for the size in bits of the type @code{char} on the
1394 target machine. If you don't define this, the default is
1395 @code{BITS_PER_UNIT}.
1397 @findex MAX_CHAR_TYPE_SIZE
1398 @item MAX_CHAR_TYPE_SIZE
1399 Maximum number for the size in bits of the type @code{char} on the
1400 target machine. If this is undefined, the default is
1401 @code{CHAR_TYPE_SIZE}. Otherwise, it is the constant value that is the
1402 largest value that @code{CHAR_TYPE_SIZE} can have at run-time. This is
1405 @findex BOOL_TYPE_SIZE
1406 @item BOOL_TYPE_SIZE
1407 A C expression for the size in bits of the C++ type @code{bool} on the
1408 target machine. If you don't define this, the default is
1409 @code{CHAR_TYPE_SIZE}.
1411 @findex FLOAT_TYPE_SIZE
1412 @item FLOAT_TYPE_SIZE
1413 A C expression for the size in bits of the type @code{float} on the
1414 target machine. If you don't define this, the default is one word.
1416 @findex DOUBLE_TYPE_SIZE
1417 @item DOUBLE_TYPE_SIZE
1418 A C expression for the size in bits of the type @code{double} on the
1419 target machine. If you don't define this, the default is two
1422 @findex LONG_DOUBLE_TYPE_SIZE
1423 @item LONG_DOUBLE_TYPE_SIZE
1424 A C expression for the size in bits of the type @code{long double} on
1425 the target machine. If you don't define this, the default is two
1428 @findex MAX_LONG_DOUBLE_TYPE_SIZE
1429 Maximum number for the size in bits of the type @code{long double} on the
1430 target machine. If this is undefined, the default is
1431 @code{LONG_DOUBLE_TYPE_SIZE}. Otherwise, it is the constant value that is
1432 the largest value that @code{LONG_DOUBLE_TYPE_SIZE} can have at run-time.
1433 This is used in @code{cpp}.
1435 @findex INTEL_EXTENDED_IEEE_FORMAT
1436 Define this macro to be 1 if the target machine uses 80-bit floating-point
1437 values with 128-bit size and alignment. This is used in @file{real.c}.
1439 @findex WIDEST_HARDWARE_FP_SIZE
1440 @item WIDEST_HARDWARE_FP_SIZE
1441 A C expression for the size in bits of the widest floating-point format
1442 supported by the hardware. If you define this macro, you must specify a
1443 value less than or equal to the value of @code{LONG_DOUBLE_TYPE_SIZE}.
1444 If you do not define this macro, the value of @code{LONG_DOUBLE_TYPE_SIZE}
1447 @findex DEFAULT_SIGNED_CHAR
1448 @item DEFAULT_SIGNED_CHAR
1449 An expression whose value is 1 or 0, according to whether the type
1450 @code{char} should be signed or unsigned by default. The user can
1451 always override this default with the options @option{-fsigned-char}
1452 and @option{-funsigned-char}.
1454 @findex DEFAULT_SHORT_ENUMS
1455 @item DEFAULT_SHORT_ENUMS
1456 A C expression to determine whether to give an @code{enum} type
1457 only as many bytes as it takes to represent the range of possible values
1458 of that type. A nonzero value means to do that; a zero value means all
1459 @code{enum} types should be allocated like @code{int}.
1461 If you don't define the macro, the default is 0.
1465 A C expression for a string describing the name of the data type to use
1466 for size values. The typedef name @code{size_t} is defined using the
1467 contents of the string.
1469 The string can contain more than one keyword. If so, separate them with
1470 spaces, and write first any length keyword, then @code{unsigned} if
1471 appropriate, and finally @code{int}. The string must exactly match one
1472 of the data type names defined in the function
1473 @code{init_decl_processing} in the file @file{c-decl.c}. You may not
1474 omit @code{int} or change the order---that would cause the compiler to
1477 If you don't define this macro, the default is @code{"long unsigned
1480 @findex PTRDIFF_TYPE
1482 A C expression for a string describing the name of the data type to use
1483 for the result of subtracting two pointers. The typedef name
1484 @code{ptrdiff_t} is defined using the contents of the string. See
1485 @code{SIZE_TYPE} above for more information.
1487 If you don't define this macro, the default is @code{"long int"}.
1491 A C expression for a string describing the name of the data type to use
1492 for wide characters. The typedef name @code{wchar_t} is defined using
1493 the contents of the string. See @code{SIZE_TYPE} above for more
1496 If you don't define this macro, the default is @code{"int"}.
1498 @findex WCHAR_TYPE_SIZE
1499 @item WCHAR_TYPE_SIZE
1500 A C expression for the size in bits of the data type for wide
1501 characters. This is used in @code{cpp}, which cannot make use of
1504 @findex MAX_WCHAR_TYPE_SIZE
1505 @item MAX_WCHAR_TYPE_SIZE
1506 Maximum number for the size in bits of the data type for wide
1507 characters. If this is undefined, the default is
1508 @code{WCHAR_TYPE_SIZE}. Otherwise, it is the constant value that is the
1509 largest value that @code{WCHAR_TYPE_SIZE} can have at run-time. This is
1512 @findex GCOV_TYPE_SIZE
1513 @item GCOV_TYPE_SIZE
1514 A C expression for the size in bits of the type used for gcov counters on the
1515 target machine. If you don't define this, the default is one
1516 @code{LONG_TYPE_SIZE} in case it is greater or equal to 64-bit and
1517 @code{LONG_LONG_TYPE_SIZE} otherwise. You may want to re-define the type to
1518 ensure atomicity for counters in multithreaded programs.
1522 A C expression for a string describing the name of the data type to
1523 use for wide characters passed to @code{printf} and returned from
1524 @code{getwc}. The typedef name @code{wint_t} is defined using the
1525 contents of the string. See @code{SIZE_TYPE} above for more
1528 If you don't define this macro, the default is @code{"unsigned int"}.
1532 A C expression for a string describing the name of the data type that
1533 can represent any value of any standard or extended signed integer type.
1534 The typedef name @code{intmax_t} is defined using the contents of the
1535 string. See @code{SIZE_TYPE} above for more information.
1537 If you don't define this macro, the default is the first of
1538 @code{"int"}, @code{"long int"}, or @code{"long long int"} that has as
1539 much precision as @code{long long int}.
1541 @findex UINTMAX_TYPE
1543 A C expression for a string describing the name of the data type that
1544 can represent any value of any standard or extended unsigned integer
1545 type. The typedef name @code{uintmax_t} is defined using the contents
1546 of the string. See @code{SIZE_TYPE} above for more information.
1548 If you don't define this macro, the default is the first of
1549 @code{"unsigned int"}, @code{"long unsigned int"}, or @code{"long long
1550 unsigned int"} that has as much precision as @code{long long unsigned
1553 @findex TARGET_PTRMEMFUNC_VBIT_LOCATION
1554 @item TARGET_PTRMEMFUNC_VBIT_LOCATION
1555 The C++ compiler represents a pointer-to-member-function with a struct
1562 ptrdiff_t vtable_index;
1569 The C++ compiler must use one bit to indicate whether the function that
1570 will be called through a pointer-to-member-function is virtual.
1571 Normally, we assume that the low-order bit of a function pointer must
1572 always be zero. Then, by ensuring that the vtable_index is odd, we can
1573 distinguish which variant of the union is in use. But, on some
1574 platforms function pointers can be odd, and so this doesn't work. In
1575 that case, we use the low-order bit of the @code{delta} field, and shift
1576 the remainder of the @code{delta} field to the left.
1578 GCC will automatically make the right selection about where to store
1579 this bit using the @code{FUNCTION_BOUNDARY} setting for your platform.
1580 However, some platforms such as ARM/Thumb have @code{FUNCTION_BOUNDARY}
1581 set such that functions always start at even addresses, but the lowest
1582 bit of pointers to functions indicate whether the function at that
1583 address is in ARM or Thumb mode. If this is the case of your
1584 architecture, you should define this macro to
1585 @code{ptrmemfunc_vbit_in_delta}.
1587 In general, you should not have to define this macro. On architectures
1588 in which function addresses are always even, according to
1589 @code{FUNCTION_BOUNDARY}, GCC will automatically define this macro to
1590 @code{ptrmemfunc_vbit_in_pfn}.
1592 @findex TARGET_VTABLE_USES_DESCRIPTORS
1593 @item TARGET_VTABLE_USES_DESCRIPTORS
1594 Normally, the C++ compiler uses function pointers in vtables. This
1595 macro allows the target to change to use ``function descriptors''
1596 instead. Function descriptors are found on targets for whom a
1597 function pointer is actually a small data structure. Normally the
1598 data structure consists of the actual code address plus a data
1599 pointer to which the function's data is relative.
1601 If vtables are used, the value of this macro should be the number
1602 of words that the function descriptor occupies.
1605 @node Escape Sequences
1606 @section Target Character Escape Sequences
1607 @cindex escape sequences
1609 By default, GCC assumes that the C character escape sequences take on
1610 their ASCII values for the target. If this is not correct, you must
1611 explicitly define all of the macros below.
1616 A C constant expression for the integer value for escape sequence
1621 A C constant expression for the integer value of the target escape
1622 character. As an extension, GCC evaluates the escape sequences
1623 @samp{\e} and @samp{\E} to this.
1627 @findex TARGET_NEWLINE
1630 @itemx TARGET_NEWLINE
1631 C constant expressions for the integer values for escape sequences
1632 @samp{\b}, @samp{\t} and @samp{\n}.
1640 C constant expressions for the integer values for escape sequences
1641 @samp{\v}, @samp{\f} and @samp{\r}.
1645 @section Register Usage
1646 @cindex register usage
1648 This section explains how to describe what registers the target machine
1649 has, and how (in general) they can be used.
1651 The description of which registers a specific instruction can use is
1652 done with register classes; see @ref{Register Classes}. For information
1653 on using registers to access a stack frame, see @ref{Frame Registers}.
1654 For passing values in registers, see @ref{Register Arguments}.
1655 For returning values in registers, see @ref{Scalar Return}.
1658 * Register Basics:: Number and kinds of registers.
1659 * Allocation Order:: Order in which registers are allocated.
1660 * Values in Registers:: What kinds of values each reg can hold.
1661 * Leaf Functions:: Renumbering registers for leaf functions.
1662 * Stack Registers:: Handling a register stack such as 80387.
1665 @node Register Basics
1666 @subsection Basic Characteristics of Registers
1668 @c prevent bad page break with this line
1669 Registers have various characteristics.
1672 @findex FIRST_PSEUDO_REGISTER
1673 @item FIRST_PSEUDO_REGISTER
1674 Number of hardware registers known to the compiler. They receive
1675 numbers 0 through @code{FIRST_PSEUDO_REGISTER-1}; thus, the first
1676 pseudo register's number really is assigned the number
1677 @code{FIRST_PSEUDO_REGISTER}.
1679 @item FIXED_REGISTERS
1680 @findex FIXED_REGISTERS
1681 @cindex fixed register
1682 An initializer that says which registers are used for fixed purposes
1683 all throughout the compiled code and are therefore not available for
1684 general allocation. These would include the stack pointer, the frame
1685 pointer (except on machines where that can be used as a general
1686 register when no frame pointer is needed), the program counter on
1687 machines where that is considered one of the addressable registers,
1688 and any other numbered register with a standard use.
1690 This information is expressed as a sequence of numbers, separated by
1691 commas and surrounded by braces. The @var{n}th number is 1 if
1692 register @var{n} is fixed, 0 otherwise.
1694 The table initialized from this macro, and the table initialized by
1695 the following one, may be overridden at run time either automatically,
1696 by the actions of the macro @code{CONDITIONAL_REGISTER_USAGE}, or by
1697 the user with the command options @option{-ffixed-@var{reg}},
1698 @option{-fcall-used-@var{reg}} and @option{-fcall-saved-@var{reg}}.
1700 @findex CALL_USED_REGISTERS
1701 @item CALL_USED_REGISTERS
1702 @cindex call-used register
1703 @cindex call-clobbered register
1704 @cindex call-saved register
1705 Like @code{FIXED_REGISTERS} but has 1 for each register that is
1706 clobbered (in general) by function calls as well as for fixed
1707 registers. This macro therefore identifies the registers that are not
1708 available for general allocation of values that must live across
1711 If a register has 0 in @code{CALL_USED_REGISTERS}, the compiler
1712 automatically saves it on function entry and restores it on function
1713 exit, if the register is used within the function.
1715 @findex CALL_REALLY_USED_REGISTERS
1716 @item CALL_REALLY_USED_REGISTERS
1717 @cindex call-used register
1718 @cindex call-clobbered register
1719 @cindex call-saved register
1720 Like @code{CALL_USED_REGISTERS} except this macro doesn't require
1721 that the entire set of @code{FIXED_REGISTERS} be included.
1722 (@code{CALL_USED_REGISTERS} must be a superset of @code{FIXED_REGISTERS}).
1723 This macro is optional. If not specified, it defaults to the value
1724 of @code{CALL_USED_REGISTERS}.
1726 @findex HARD_REGNO_CALL_PART_CLOBBERED
1727 @item HARD_REGNO_CALL_PART_CLOBBERED (@var{regno}, @var{mode})
1728 @cindex call-used register
1729 @cindex call-clobbered register
1730 @cindex call-saved register
1731 A C expression that is nonzero if it is not permissible to store a
1732 value of mode @var{mode} in hard register number @var{regno} across a
1733 call without some part of it being clobbered. For most machines this
1734 macro need not be defined. It is only required for machines that do not
1735 preserve the entire contents of a register across a call.
1737 @findex CONDITIONAL_REGISTER_USAGE
1739 @findex call_used_regs
1740 @item CONDITIONAL_REGISTER_USAGE
1741 Zero or more C statements that may conditionally modify five variables
1742 @code{fixed_regs}, @code{call_used_regs}, @code{global_regs},
1743 @code{reg_names}, and @code{reg_class_contents}, to take into account
1744 any dependence of these register sets on target flags. The first three
1745 of these are of type @code{char []} (interpreted as Boolean vectors).
1746 @code{global_regs} is a @code{const char *[]}, and
1747 @code{reg_class_contents} is a @code{HARD_REG_SET}. Before the macro is
1748 called, @code{fixed_regs}, @code{call_used_regs},
1749 @code{reg_class_contents}, and @code{reg_names} have been initialized
1750 from @code{FIXED_REGISTERS}, @code{CALL_USED_REGISTERS},
1751 @code{REG_CLASS_CONTENTS}, and @code{REGISTER_NAMES}, respectively.
1752 @code{global_regs} has been cleared, and any @option{-ffixed-@var{reg}},
1753 @option{-fcall-used-@var{reg}} and @option{-fcall-saved-@var{reg}}
1754 command options have been applied.
1756 You need not define this macro if it has no work to do.
1758 @cindex disabling certain registers
1759 @cindex controlling register usage
1760 If the usage of an entire class of registers depends on the target
1761 flags, you may indicate this to GCC by using this macro to modify
1762 @code{fixed_regs} and @code{call_used_regs} to 1 for each of the
1763 registers in the classes which should not be used by GCC@. Also define
1764 the macro @code{REG_CLASS_FROM_LETTER} to return @code{NO_REGS} if it
1765 is called with a letter for a class that shouldn't be used.
1767 (However, if this class is not included in @code{GENERAL_REGS} and all
1768 of the insn patterns whose constraints permit this class are
1769 controlled by target switches, then GCC will automatically avoid using
1770 these registers when the target switches are opposed to them.)
1772 @findex NON_SAVING_SETJMP
1773 @item NON_SAVING_SETJMP
1774 If this macro is defined and has a nonzero value, it means that
1775 @code{setjmp} and related functions fail to save the registers, or that
1776 @code{longjmp} fails to restore them. To compensate, the compiler
1777 avoids putting variables in registers in functions that use
1780 @findex INCOMING_REGNO
1781 @item INCOMING_REGNO (@var{out})
1782 Define this macro if the target machine has register windows. This C
1783 expression returns the register number as seen by the called function
1784 corresponding to the register number @var{out} as seen by the calling
1785 function. Return @var{out} if register number @var{out} is not an
1788 @findex OUTGOING_REGNO
1789 @item OUTGOING_REGNO (@var{in})
1790 Define this macro if the target machine has register windows. This C
1791 expression returns the register number as seen by the calling function
1792 corresponding to the register number @var{in} as seen by the called
1793 function. Return @var{in} if register number @var{in} is not an inbound
1797 @item LOCAL_REGNO (@var{regno})
1798 Define this macro if the target machine has register windows. This C
1799 expression returns true if the register is call-saved but is in the
1800 register window. Unlike most call-saved registers, such registers
1801 need not be explicitly restored on function exit or during non-local
1807 If the program counter has a register number, define this as that
1808 register number. Otherwise, do not define it.
1812 @node Allocation Order
1813 @subsection Order of Allocation of Registers
1814 @cindex order of register allocation
1815 @cindex register allocation order
1817 @c prevent bad page break with this line
1818 Registers are allocated in order.
1821 @findex REG_ALLOC_ORDER
1822 @item REG_ALLOC_ORDER
1823 If defined, an initializer for a vector of integers, containing the
1824 numbers of hard registers in the order in which GCC should prefer
1825 to use them (from most preferred to least).
1827 If this macro is not defined, registers are used lowest numbered first
1828 (all else being equal).
1830 One use of this macro is on machines where the highest numbered
1831 registers must always be saved and the save-multiple-registers
1832 instruction supports only sequences of consecutive registers. On such
1833 machines, define @code{REG_ALLOC_ORDER} to be an initializer that lists
1834 the highest numbered allocable register first.
1836 @findex ORDER_REGS_FOR_LOCAL_ALLOC
1837 @item ORDER_REGS_FOR_LOCAL_ALLOC
1838 A C statement (sans semicolon) to choose the order in which to allocate
1839 hard registers for pseudo-registers local to a basic block.
1841 Store the desired register order in the array @code{reg_alloc_order}.
1842 Element 0 should be the register to allocate first; element 1, the next
1843 register; and so on.
1845 The macro body should not assume anything about the contents of
1846 @code{reg_alloc_order} before execution of the macro.
1848 On most machines, it is not necessary to define this macro.
1851 @node Values in Registers
1852 @subsection How Values Fit in Registers
1854 This section discusses the macros that describe which kinds of values
1855 (specifically, which machine modes) each register can hold, and how many
1856 consecutive registers are needed for a given mode.
1859 @findex HARD_REGNO_NREGS
1860 @item HARD_REGNO_NREGS (@var{regno}, @var{mode})
1861 A C expression for the number of consecutive hard registers, starting
1862 at register number @var{regno}, required to hold a value of mode
1865 On a machine where all registers are exactly one word, a suitable
1866 definition of this macro is
1869 #define HARD_REGNO_NREGS(REGNO, MODE) \
1870 ((GET_MODE_SIZE (MODE) + UNITS_PER_WORD - 1) \
1874 @findex HARD_REGNO_MODE_OK
1875 @item HARD_REGNO_MODE_OK (@var{regno}, @var{mode})
1876 A C expression that is nonzero if it is permissible to store a value
1877 of mode @var{mode} in hard register number @var{regno} (or in several
1878 registers starting with that one). For a machine where all registers
1879 are equivalent, a suitable definition is
1882 #define HARD_REGNO_MODE_OK(REGNO, MODE) 1
1885 You need not include code to check for the numbers of fixed registers,
1886 because the allocation mechanism considers them to be always occupied.
1888 @cindex register pairs
1889 On some machines, double-precision values must be kept in even/odd
1890 register pairs. You can implement that by defining this macro to reject
1891 odd register numbers for such modes.
1893 The minimum requirement for a mode to be OK in a register is that the
1894 @samp{mov@var{mode}} instruction pattern support moves between the
1895 register and other hard register in the same class and that moving a
1896 value into the register and back out not alter it.
1898 Since the same instruction used to move @code{word_mode} will work for
1899 all narrower integer modes, it is not necessary on any machine for
1900 @code{HARD_REGNO_MODE_OK} to distinguish between these modes, provided
1901 you define patterns @samp{movhi}, etc., to take advantage of this. This
1902 is useful because of the interaction between @code{HARD_REGNO_MODE_OK}
1903 and @code{MODES_TIEABLE_P}; it is very desirable for all integer modes
1906 Many machines have special registers for floating point arithmetic.
1907 Often people assume that floating point machine modes are allowed only
1908 in floating point registers. This is not true. Any registers that
1909 can hold integers can safely @emph{hold} a floating point machine
1910 mode, whether or not floating arithmetic can be done on it in those
1911 registers. Integer move instructions can be used to move the values.
1913 On some machines, though, the converse is true: fixed-point machine
1914 modes may not go in floating registers. This is true if the floating
1915 registers normalize any value stored in them, because storing a
1916 non-floating value there would garble it. In this case,
1917 @code{HARD_REGNO_MODE_OK} should reject fixed-point machine modes in
1918 floating registers. But if the floating registers do not automatically
1919 normalize, if you can store any bit pattern in one and retrieve it
1920 unchanged without a trap, then any machine mode may go in a floating
1921 register, so you can define this macro to say so.
1923 The primary significance of special floating registers is rather that
1924 they are the registers acceptable in floating point arithmetic
1925 instructions. However, this is of no concern to
1926 @code{HARD_REGNO_MODE_OK}. You handle it by writing the proper
1927 constraints for those instructions.
1929 On some machines, the floating registers are especially slow to access,
1930 so that it is better to store a value in a stack frame than in such a
1931 register if floating point arithmetic is not being done. As long as the
1932 floating registers are not in class @code{GENERAL_REGS}, they will not
1933 be used unless some pattern's constraint asks for one.
1935 @findex MODES_TIEABLE_P
1936 @item MODES_TIEABLE_P (@var{mode1}, @var{mode2})
1937 A C expression that is nonzero if a value of mode
1938 @var{mode1} is accessible in mode @var{mode2} without copying.
1940 If @code{HARD_REGNO_MODE_OK (@var{r}, @var{mode1})} and
1941 @code{HARD_REGNO_MODE_OK (@var{r}, @var{mode2})} are always the same for
1942 any @var{r}, then @code{MODES_TIEABLE_P (@var{mode1}, @var{mode2})}
1943 should be nonzero. If they differ for any @var{r}, you should define
1944 this macro to return zero unless some other mechanism ensures the
1945 accessibility of the value in a narrower mode.
1947 You should define this macro to return nonzero in as many cases as
1948 possible since doing so will allow GCC to perform better register
1951 @findex AVOID_CCMODE_COPIES
1952 @item AVOID_CCMODE_COPIES
1953 Define this macro if the compiler should avoid copies to/from @code{CCmode}
1954 registers. You should only define this macro if support for copying to/from
1955 @code{CCmode} is incomplete.
1957 @findex SUBREG_REGNO_OFFSET
1958 @item SUBREG_REGNO_OFFSET
1959 Define this macro if the compiler needs to handle subregs in a non-standard
1960 way. The macro returns the correct regno offset for mode @code{YMODE} given
1961 a subreg of type @code{XMODE}.
1962 This macro takes 4 parameters:
1965 A regno of an inner hard subreg_reg (or what will become one).
1971 The mode of a top level SUBREG (or what may become one).
1973 The default function can be found in @file{rtlanal.c}, function
1974 @code{subreg_regno_offset}. Normally this does not need to be defined.
1977 @node Leaf Functions
1978 @subsection Handling Leaf Functions
1980 @cindex leaf functions
1981 @cindex functions, leaf
1982 On some machines, a leaf function (i.e., one which makes no calls) can run
1983 more efficiently if it does not make its own register window. Often this
1984 means it is required to receive its arguments in the registers where they
1985 are passed by the caller, instead of the registers where they would
1988 The special treatment for leaf functions generally applies only when
1989 other conditions are met; for example, often they may use only those
1990 registers for its own variables and temporaries. We use the term ``leaf
1991 function'' to mean a function that is suitable for this special
1992 handling, so that functions with no calls are not necessarily ``leaf
1995 GCC assigns register numbers before it knows whether the function is
1996 suitable for leaf function treatment. So it needs to renumber the
1997 registers in order to output a leaf function. The following macros
2001 @findex LEAF_REGISTERS
2002 @item LEAF_REGISTERS
2003 Name of a char vector, indexed by hard register number, which
2004 contains 1 for a register that is allowable in a candidate for leaf
2007 If leaf function treatment involves renumbering the registers, then the
2008 registers marked here should be the ones before renumbering---those that
2009 GCC would ordinarily allocate. The registers which will actually be
2010 used in the assembler code, after renumbering, should not be marked with 1
2013 Define this macro only if the target machine offers a way to optimize
2014 the treatment of leaf functions.
2016 @findex LEAF_REG_REMAP
2017 @item LEAF_REG_REMAP (@var{regno})
2018 A C expression whose value is the register number to which @var{regno}
2019 should be renumbered, when a function is treated as a leaf function.
2021 If @var{regno} is a register number which should not appear in a leaf
2022 function before renumbering, then the expression should yield @minus{}1, which
2023 will cause the compiler to abort.
2025 Define this macro only if the target machine offers a way to optimize the
2026 treatment of leaf functions, and registers need to be renumbered to do
2030 @findex current_function_is_leaf
2031 @findex current_function_uses_only_leaf_regs
2032 @code{TARGET_ASM_FUNCTION_PROLOGUE} and
2033 @code{TARGET_ASM_FUNCTION_EPILOGUE} must usually treat leaf functions
2034 specially. They can test the C variable @code{current_function_is_leaf}
2035 which is nonzero for leaf functions. @code{current_function_is_leaf} is
2036 set prior to local register allocation and is valid for the remaining
2037 compiler passes. They can also test the C variable
2038 @code{current_function_uses_only_leaf_regs} which is nonzero for leaf
2039 functions which only use leaf registers.
2040 @code{current_function_uses_only_leaf_regs} is valid after reload and is
2041 only useful if @code{LEAF_REGISTERS} is defined.
2042 @c changed this to fix overfull. ALSO: why the "it" at the beginning
2043 @c of the next paragraph?! --mew 2feb93
2045 @node Stack Registers
2046 @subsection Registers That Form a Stack
2048 There are special features to handle computers where some of the
2049 ``registers'' form a stack, as in the 80387 coprocessor for the 80386.
2050 Stack registers are normally written by pushing onto the stack, and are
2051 numbered relative to the top of the stack.
2053 Currently, GCC can only handle one group of stack-like registers, and
2054 they must be consecutively numbered.
2059 Define this if the machine has any stack-like registers.
2061 @findex FIRST_STACK_REG
2062 @item FIRST_STACK_REG
2063 The number of the first stack-like register. This one is the top
2066 @findex LAST_STACK_REG
2067 @item LAST_STACK_REG
2068 The number of the last stack-like register. This one is the bottom of
2072 @node Register Classes
2073 @section Register Classes
2074 @cindex register class definitions
2075 @cindex class definitions, register
2077 On many machines, the numbered registers are not all equivalent.
2078 For example, certain registers may not be allowed for indexed addressing;
2079 certain registers may not be allowed in some instructions. These machine
2080 restrictions are described to the compiler using @dfn{register classes}.
2082 You define a number of register classes, giving each one a name and saying
2083 which of the registers belong to it. Then you can specify register classes
2084 that are allowed as operands to particular instruction patterns.
2088 In general, each register will belong to several classes. In fact, one
2089 class must be named @code{ALL_REGS} and contain all the registers. Another
2090 class must be named @code{NO_REGS} and contain no registers. Often the
2091 union of two classes will be another class; however, this is not required.
2093 @findex GENERAL_REGS
2094 One of the classes must be named @code{GENERAL_REGS}. There is nothing
2095 terribly special about the name, but the operand constraint letters
2096 @samp{r} and @samp{g} specify this class. If @code{GENERAL_REGS} is
2097 the same as @code{ALL_REGS}, just define it as a macro which expands
2100 Order the classes so that if class @var{x} is contained in class @var{y}
2101 then @var{x} has a lower class number than @var{y}.
2103 The way classes other than @code{GENERAL_REGS} are specified in operand
2104 constraints is through machine-dependent operand constraint letters.
2105 You can define such letters to correspond to various classes, then use
2106 them in operand constraints.
2108 You should define a class for the union of two classes whenever some
2109 instruction allows both classes. For example, if an instruction allows
2110 either a floating point (coprocessor) register or a general register for a
2111 certain operand, you should define a class @code{FLOAT_OR_GENERAL_REGS}
2112 which includes both of them. Otherwise you will get suboptimal code.
2114 You must also specify certain redundant information about the register
2115 classes: for each class, which classes contain it and which ones are
2116 contained in it; for each pair of classes, the largest class contained
2119 When a value occupying several consecutive registers is expected in a
2120 certain class, all the registers used must belong to that class.
2121 Therefore, register classes cannot be used to enforce a requirement for
2122 a register pair to start with an even-numbered register. The way to
2123 specify this requirement is with @code{HARD_REGNO_MODE_OK}.
2125 Register classes used for input-operands of bitwise-and or shift
2126 instructions have a special requirement: each such class must have, for
2127 each fixed-point machine mode, a subclass whose registers can transfer that
2128 mode to or from memory. For example, on some machines, the operations for
2129 single-byte values (@code{QImode}) are limited to certain registers. When
2130 this is so, each register class that is used in a bitwise-and or shift
2131 instruction must have a subclass consisting of registers from which
2132 single-byte values can be loaded or stored. This is so that
2133 @code{PREFERRED_RELOAD_CLASS} can always have a possible value to return.
2136 @findex enum reg_class
2137 @item enum reg_class
2138 An enumeral type that must be defined with all the register class names
2139 as enumeral values. @code{NO_REGS} must be first. @code{ALL_REGS}
2140 must be the last register class, followed by one more enumeral value,
2141 @code{LIM_REG_CLASSES}, which is not a register class but rather
2142 tells how many classes there are.
2144 Each register class has a number, which is the value of casting
2145 the class name to type @code{int}. The number serves as an index
2146 in many of the tables described below.
2148 @findex N_REG_CLASSES
2150 The number of distinct register classes, defined as follows:
2153 #define N_REG_CLASSES (int) LIM_REG_CLASSES
2156 @findex REG_CLASS_NAMES
2157 @item REG_CLASS_NAMES
2158 An initializer containing the names of the register classes as C string
2159 constants. These names are used in writing some of the debugging dumps.
2161 @findex REG_CLASS_CONTENTS
2162 @item REG_CLASS_CONTENTS
2163 An initializer containing the contents of the register classes, as integers
2164 which are bit masks. The @var{n}th integer specifies the contents of class
2165 @var{n}. The way the integer @var{mask} is interpreted is that
2166 register @var{r} is in the class if @code{@var{mask} & (1 << @var{r})} is 1.
2168 When the machine has more than 32 registers, an integer does not suffice.
2169 Then the integers are replaced by sub-initializers, braced groupings containing
2170 several integers. Each sub-initializer must be suitable as an initializer
2171 for the type @code{HARD_REG_SET} which is defined in @file{hard-reg-set.h}.
2172 In this situation, the first integer in each sub-initializer corresponds to
2173 registers 0 through 31, the second integer to registers 32 through 63, and
2176 @findex REGNO_REG_CLASS
2177 @item REGNO_REG_CLASS (@var{regno})
2178 A C expression whose value is a register class containing hard register
2179 @var{regno}. In general there is more than one such class; choose a class
2180 which is @dfn{minimal}, meaning that no smaller class also contains the
2183 @findex BASE_REG_CLASS
2184 @item BASE_REG_CLASS
2185 A macro whose definition is the name of the class to which a valid
2186 base register must belong. A base register is one used in an address
2187 which is the register value plus a displacement.
2189 @findex MODE_BASE_REG_CLASS
2190 @item MODE_BASE_REG_CLASS (@var{mode})
2191 This is a variation of the @code{BASE_REG_CLASS} macro which allows
2192 the selection of a base register in a mode depenedent manner. If
2193 @var{mode} is VOIDmode then it should return the same value as
2194 @code{BASE_REG_CLASS}.
2196 @findex INDEX_REG_CLASS
2197 @item INDEX_REG_CLASS
2198 A macro whose definition is the name of the class to which a valid
2199 index register must belong. An index register is one used in an
2200 address where its value is either multiplied by a scale factor or
2201 added to another register (as well as added to a displacement).
2203 @findex REG_CLASS_FROM_LETTER
2204 @item REG_CLASS_FROM_LETTER (@var{char})
2205 A C expression which defines the machine-dependent operand constraint
2206 letters for register classes. If @var{char} is such a letter, the
2207 value should be the register class corresponding to it. Otherwise,
2208 the value should be @code{NO_REGS}. The register letter @samp{r},
2209 corresponding to class @code{GENERAL_REGS}, will not be passed
2210 to this macro; you do not need to handle it.
2212 @findex REGNO_OK_FOR_BASE_P
2213 @item REGNO_OK_FOR_BASE_P (@var{num})
2214 A C expression which is nonzero if register number @var{num} is
2215 suitable for use as a base register in operand addresses. It may be
2216 either a suitable hard register or a pseudo register that has been
2217 allocated such a hard register.
2219 @findex REGNO_MODE_OK_FOR_BASE_P
2220 @item REGNO_MODE_OK_FOR_BASE_P (@var{num}, @var{mode})
2221 A C expression that is just like @code{REGNO_OK_FOR_BASE_P}, except that
2222 that expression may examine the mode of the memory reference in
2223 @var{mode}. You should define this macro if the mode of the memory
2224 reference affects whether a register may be used as a base register. If
2225 you define this macro, the compiler will use it instead of
2226 @code{REGNO_OK_FOR_BASE_P}.
2228 @findex REGNO_OK_FOR_INDEX_P
2229 @item REGNO_OK_FOR_INDEX_P (@var{num})
2230 A C expression which is nonzero if register number @var{num} is
2231 suitable for use as an index register in operand addresses. It may be
2232 either a suitable hard register or a pseudo register that has been
2233 allocated such a hard register.
2235 The difference between an index register and a base register is that
2236 the index register may be scaled. If an address involves the sum of
2237 two registers, neither one of them scaled, then either one may be
2238 labeled the ``base'' and the other the ``index''; but whichever
2239 labeling is used must fit the machine's constraints of which registers
2240 may serve in each capacity. The compiler will try both labelings,
2241 looking for one that is valid, and will reload one or both registers
2242 only if neither labeling works.
2244 @findex PREFERRED_RELOAD_CLASS
2245 @item PREFERRED_RELOAD_CLASS (@var{x}, @var{class})
2246 A C expression that places additional restrictions on the register class
2247 to use when it is necessary to copy value @var{x} into a register in class
2248 @var{class}. The value is a register class; perhaps @var{class}, or perhaps
2249 another, smaller class. On many machines, the following definition is
2253 #define PREFERRED_RELOAD_CLASS(X,CLASS) CLASS
2256 Sometimes returning a more restrictive class makes better code. For
2257 example, on the 68000, when @var{x} is an integer constant that is in range
2258 for a @samp{moveq} instruction, the value of this macro is always
2259 @code{DATA_REGS} as long as @var{class} includes the data registers.
2260 Requiring a data register guarantees that a @samp{moveq} will be used.
2262 If @var{x} is a @code{const_double}, by returning @code{NO_REGS}
2263 you can force @var{x} into a memory constant. This is useful on
2264 certain machines where immediate floating values cannot be loaded into
2265 certain kinds of registers.
2267 @findex PREFERRED_OUTPUT_RELOAD_CLASS
2268 @item PREFERRED_OUTPUT_RELOAD_CLASS (@var{x}, @var{class})
2269 Like @code{PREFERRED_RELOAD_CLASS}, but for output reloads instead of
2270 input reloads. If you don't define this macro, the default is to use
2271 @var{class}, unchanged.
2273 @findex LIMIT_RELOAD_CLASS
2274 @item LIMIT_RELOAD_CLASS (@var{mode}, @var{class})
2275 A C expression that places additional restrictions on the register class
2276 to use when it is necessary to be able to hold a value of mode
2277 @var{mode} in a reload register for which class @var{class} would
2280 Unlike @code{PREFERRED_RELOAD_CLASS}, this macro should be used when
2281 there are certain modes that simply can't go in certain reload classes.
2283 The value is a register class; perhaps @var{class}, or perhaps another,
2286 Don't define this macro unless the target machine has limitations which
2287 require the macro to do something nontrivial.
2289 @findex SECONDARY_RELOAD_CLASS
2290 @findex SECONDARY_INPUT_RELOAD_CLASS
2291 @findex SECONDARY_OUTPUT_RELOAD_CLASS
2292 @item SECONDARY_RELOAD_CLASS (@var{class}, @var{mode}, @var{x})
2293 @itemx SECONDARY_INPUT_RELOAD_CLASS (@var{class}, @var{mode}, @var{x})
2294 @itemx SECONDARY_OUTPUT_RELOAD_CLASS (@var{class}, @var{mode}, @var{x})
2295 Many machines have some registers that cannot be copied directly to or
2296 from memory or even from other types of registers. An example is the
2297 @samp{MQ} register, which on most machines, can only be copied to or
2298 from general registers, but not memory. Some machines allow copying all
2299 registers to and from memory, but require a scratch register for stores
2300 to some memory locations (e.g., those with symbolic address on the RT,
2301 and those with certain symbolic address on the Sparc when compiling
2302 PIC)@. In some cases, both an intermediate and a scratch register are
2305 You should define these macros to indicate to the reload phase that it may
2306 need to allocate at least one register for a reload in addition to the
2307 register to contain the data. Specifically, if copying @var{x} to a
2308 register @var{class} in @var{mode} requires an intermediate register,
2309 you should define @code{SECONDARY_INPUT_RELOAD_CLASS} to return the
2310 largest register class all of whose registers can be used as
2311 intermediate registers or scratch registers.
2313 If copying a register @var{class} in @var{mode} to @var{x} requires an
2314 intermediate or scratch register, @code{SECONDARY_OUTPUT_RELOAD_CLASS}
2315 should be defined to return the largest register class required. If the
2316 requirements for input and output reloads are the same, the macro
2317 @code{SECONDARY_RELOAD_CLASS} should be used instead of defining both
2320 The values returned by these macros are often @code{GENERAL_REGS}.
2321 Return @code{NO_REGS} if no spare register is needed; i.e., if @var{x}
2322 can be directly copied to or from a register of @var{class} in
2323 @var{mode} without requiring a scratch register. Do not define this
2324 macro if it would always return @code{NO_REGS}.
2326 If a scratch register is required (either with or without an
2327 intermediate register), you should define patterns for
2328 @samp{reload_in@var{m}} or @samp{reload_out@var{m}}, as required
2329 (@pxref{Standard Names}. These patterns, which will normally be
2330 implemented with a @code{define_expand}, should be similar to the
2331 @samp{mov@var{m}} patterns, except that operand 2 is the scratch
2334 Define constraints for the reload register and scratch register that
2335 contain a single register class. If the original reload register (whose
2336 class is @var{class}) can meet the constraint given in the pattern, the
2337 value returned by these macros is used for the class of the scratch
2338 register. Otherwise, two additional reload registers are required.
2339 Their classes are obtained from the constraints in the insn pattern.
2341 @var{x} might be a pseudo-register or a @code{subreg} of a
2342 pseudo-register, which could either be in a hard register or in memory.
2343 Use @code{true_regnum} to find out; it will return @minus{}1 if the pseudo is
2344 in memory and the hard register number if it is in a register.
2346 These macros should not be used in the case where a particular class of
2347 registers can only be copied to memory and not to another class of
2348 registers. In that case, secondary reload registers are not needed and
2349 would not be helpful. Instead, a stack location must be used to perform
2350 the copy and the @code{mov@var{m}} pattern should use memory as an
2351 intermediate storage. This case often occurs between floating-point and
2354 @findex SECONDARY_MEMORY_NEEDED
2355 @item SECONDARY_MEMORY_NEEDED (@var{class1}, @var{class2}, @var{m})
2356 Certain machines have the property that some registers cannot be copied
2357 to some other registers without using memory. Define this macro on
2358 those machines to be a C expression that is nonzero if objects of mode
2359 @var{m} in registers of @var{class1} can only be copied to registers of
2360 class @var{class2} by storing a register of @var{class1} into memory
2361 and loading that memory location into a register of @var{class2}.
2363 Do not define this macro if its value would always be zero.
2365 @findex SECONDARY_MEMORY_NEEDED_RTX
2366 @item SECONDARY_MEMORY_NEEDED_RTX (@var{mode})
2367 Normally when @code{SECONDARY_MEMORY_NEEDED} is defined, the compiler
2368 allocates a stack slot for a memory location needed for register copies.
2369 If this macro is defined, the compiler instead uses the memory location
2370 defined by this macro.
2372 Do not define this macro if you do not define
2373 @code{SECONDARY_MEMORY_NEEDED}.
2375 @findex SECONDARY_MEMORY_NEEDED_MODE
2376 @item SECONDARY_MEMORY_NEEDED_MODE (@var{mode})
2377 When the compiler needs a secondary memory location to copy between two
2378 registers of mode @var{mode}, it normally allocates sufficient memory to
2379 hold a quantity of @code{BITS_PER_WORD} bits and performs the store and
2380 load operations in a mode that many bits wide and whose class is the
2381 same as that of @var{mode}.
2383 This is right thing to do on most machines because it ensures that all
2384 bits of the register are copied and prevents accesses to the registers
2385 in a narrower mode, which some machines prohibit for floating-point
2388 However, this default behavior is not correct on some machines, such as
2389 the DEC Alpha, that store short integers in floating-point registers
2390 differently than in integer registers. On those machines, the default
2391 widening will not work correctly and you must define this macro to
2392 suppress that widening in some cases. See the file @file{alpha.h} for
2395 Do not define this macro if you do not define
2396 @code{SECONDARY_MEMORY_NEEDED} or if widening @var{mode} to a mode that
2397 is @code{BITS_PER_WORD} bits wide is correct for your machine.
2399 @findex SMALL_REGISTER_CLASSES
2400 @item SMALL_REGISTER_CLASSES
2401 On some machines, it is risky to let hard registers live across arbitrary
2402 insns. Typically, these machines have instructions that require values
2403 to be in specific registers (like an accumulator), and reload will fail
2404 if the required hard register is used for another purpose across such an
2407 Define @code{SMALL_REGISTER_CLASSES} to be an expression with a nonzero
2408 value on these machines. When this macro has a nonzero value, the
2409 compiler will try to minimize the lifetime of hard registers.
2411 It is always safe to define this macro with a nonzero value, but if you
2412 unnecessarily define it, you will reduce the amount of optimizations
2413 that can be performed in some cases. If you do not define this macro
2414 with a nonzero value when it is required, the compiler will run out of
2415 spill registers and print a fatal error message. For most machines, you
2416 should not define this macro at all.
2418 @findex CLASS_LIKELY_SPILLED_P
2419 @item CLASS_LIKELY_SPILLED_P (@var{class})
2420 A C expression whose value is nonzero if pseudos that have been assigned
2421 to registers of class @var{class} would likely be spilled because
2422 registers of @var{class} are needed for spill registers.
2424 The default value of this macro returns 1 if @var{class} has exactly one
2425 register and zero otherwise. On most machines, this default should be
2426 used. Only define this macro to some other expression if pseudos
2427 allocated by @file{local-alloc.c} end up in memory because their hard
2428 registers were needed for spill registers. If this macro returns nonzero
2429 for those classes, those pseudos will only be allocated by
2430 @file{global.c}, which knows how to reallocate the pseudo to another
2431 register. If there would not be another register available for
2432 reallocation, you should not change the definition of this macro since
2433 the only effect of such a definition would be to slow down register
2436 @findex CLASS_MAX_NREGS
2437 @item CLASS_MAX_NREGS (@var{class}, @var{mode})
2438 A C expression for the maximum number of consecutive registers
2439 of class @var{class} needed to hold a value of mode @var{mode}.
2441 This is closely related to the macro @code{HARD_REGNO_NREGS}. In fact,
2442 the value of the macro @code{CLASS_MAX_NREGS (@var{class}, @var{mode})}
2443 should be the maximum value of @code{HARD_REGNO_NREGS (@var{regno},
2444 @var{mode})} for all @var{regno} values in the class @var{class}.
2446 This macro helps control the handling of multiple-word values
2449 @item CLASS_CANNOT_CHANGE_MODE
2450 If defined, a C expression for a class that contains registers for
2451 which the compiler may not change modes arbitrarily.
2453 @item CLASS_CANNOT_CHANGE_MODE_P(@var{from}, @var{to})
2454 A C expression that is true if, for a register in
2455 @code{CLASS_CANNOT_CHANGE_MODE}, the requested mode punning is invalid.
2457 For the example, loading 32-bit integer or floating-point objects into
2458 floating-point registers on the Alpha extends them to 64-bits.
2459 Therefore loading a 64-bit object and then storing it as a 32-bit object
2460 does not store the low-order 32-bits, as would be the case for a normal
2461 register. Therefore, @file{alpha.h} defines @code{CLASS_CANNOT_CHANGE_MODE}
2462 as @code{FLOAT_REGS} and @code{CLASS_CANNOT_CHANGE_MODE_P} restricts
2463 mode changes to same-size modes.
2465 Compare this to IA-64, which extends floating-point values to 82-bits,
2466 and stores 64-bit integers in a different format than 64-bit doubles.
2467 Therefore @code{CLASS_CANNOT_CHANGE_MODE_P} is always true.
2470 Three other special macros describe which operands fit which constraint
2474 @findex CONST_OK_FOR_LETTER_P
2475 @item CONST_OK_FOR_LETTER_P (@var{value}, @var{c})
2476 A C expression that defines the machine-dependent operand constraint
2477 letters (@samp{I}, @samp{J}, @samp{K}, @dots{} @samp{P}) that specify
2478 particular ranges of integer values. If @var{c} is one of those
2479 letters, the expression should check that @var{value}, an integer, is in
2480 the appropriate range and return 1 if so, 0 otherwise. If @var{c} is
2481 not one of those letters, the value should be 0 regardless of
2484 @findex CONST_DOUBLE_OK_FOR_LETTER_P
2485 @item CONST_DOUBLE_OK_FOR_LETTER_P (@var{value}, @var{c})
2486 A C expression that defines the machine-dependent operand constraint
2487 letters that specify particular ranges of @code{const_double} values
2488 (@samp{G} or @samp{H}).
2490 If @var{c} is one of those letters, the expression should check that
2491 @var{value}, an RTX of code @code{const_double}, is in the appropriate
2492 range and return 1 if so, 0 otherwise. If @var{c} is not one of those
2493 letters, the value should be 0 regardless of @var{value}.
2495 @code{const_double} is used for all floating-point constants and for
2496 @code{DImode} fixed-point constants. A given letter can accept either
2497 or both kinds of values. It can use @code{GET_MODE} to distinguish
2498 between these kinds.
2500 @findex EXTRA_CONSTRAINT
2501 @item EXTRA_CONSTRAINT (@var{value}, @var{c})
2502 A C expression that defines the optional machine-dependent constraint
2503 letters that can be used to segregate specific types of operands, usually
2504 memory references, for the target machine. Any letter that is not
2505 elsewhere defined and not matched by @code{REG_CLASS_FROM_LETTER}
2506 may be used. Normally this macro will not be defined.
2508 If it is required for a particular target machine, it should return 1
2509 if @var{value} corresponds to the operand type represented by the
2510 constraint letter @var{c}. If @var{c} is not defined as an extra
2511 constraint, the value returned should be 0 regardless of @var{value}.
2513 For example, on the ROMP, load instructions cannot have their output
2514 in r0 if the memory reference contains a symbolic address. Constraint
2515 letter @samp{Q} is defined as representing a memory address that does
2516 @emph{not} contain a symbolic address. An alternative is specified with
2517 a @samp{Q} constraint on the input and @samp{r} on the output. The next
2518 alternative specifies @samp{m} on the input and a register class that
2519 does not include r0 on the output.
2522 @node Stack and Calling
2523 @section Stack Layout and Calling Conventions
2524 @cindex calling conventions
2526 @c prevent bad page break with this line
2527 This describes the stack layout and calling conventions.
2531 * Exception Handling::
2536 * Register Arguments::
2538 * Aggregate Return::
2546 @subsection Basic Stack Layout
2547 @cindex stack frame layout
2548 @cindex frame layout
2550 @c prevent bad page break with this line
2551 Here is the basic stack layout.
2554 @findex STACK_GROWS_DOWNWARD
2555 @item STACK_GROWS_DOWNWARD
2556 Define this macro if pushing a word onto the stack moves the stack
2557 pointer to a smaller address.
2559 When we say, ``define this macro if @dots{},'' it means that the
2560 compiler checks this macro only with @code{#ifdef} so the precise
2561 definition used does not matter.
2563 @findex STACK_PUSH_CODE
2564 @item STACK_PUSH_CODE
2566 This macro defines the operation used when something is pushed
2567 on the stack. In RTL, a push operation will be
2568 @code{(set (mem (STACK_PUSH_CODE (reg sp))) ...)}
2570 The choices are @code{PRE_DEC}, @code{POST_DEC}, @code{PRE_INC},
2571 and @code{POST_INC}. Which of these is correct depends on
2572 the stack direction and on whether the stack pointer points
2573 to the last item on the stack or whether it points to the
2574 space for the next item on the stack.
2576 The default is @code{PRE_DEC} when @code{STACK_GROWS_DOWNWARD} is
2577 defined, which is almost always right, and @code{PRE_INC} otherwise,
2578 which is often wrong.
2580 @findex FRAME_GROWS_DOWNWARD
2581 @item FRAME_GROWS_DOWNWARD
2582 Define this macro if the addresses of local variable slots are at negative
2583 offsets from the frame pointer.
2585 @findex ARGS_GROW_DOWNWARD
2586 @item ARGS_GROW_DOWNWARD
2587 Define this macro if successive arguments to a function occupy decreasing
2588 addresses on the stack.
2590 @findex STARTING_FRAME_OFFSET
2591 @item STARTING_FRAME_OFFSET
2592 Offset from the frame pointer to the first local variable slot to be allocated.
2594 If @code{FRAME_GROWS_DOWNWARD}, find the next slot's offset by
2595 subtracting the first slot's length from @code{STARTING_FRAME_OFFSET}.
2596 Otherwise, it is found by adding the length of the first slot to the
2597 value @code{STARTING_FRAME_OFFSET}.
2598 @c i'm not sure if the above is still correct.. had to change it to get
2599 @c rid of an overfull. --mew 2feb93
2601 @findex STARTING_FRAME_PHASE
2602 @item STARTING_FRAME_PHASE
2603 This option species how many bytes the frame is out of phase from the
2606 For example, some ports assume a stack alignment of 128 bits, but the
2607 start of the frame is 64 bits displaced from this alignment. In this
2608 case, you would define @code{STARTING_FRAME_PHASE} to be 8.
2610 This macro defaults to 0, so there is no need to define it if the start
2611 of the frame maintains the stack alignment.
2613 @findex STACK_POINTER_OFFSET
2614 @item STACK_POINTER_OFFSET
2615 Offset from the stack pointer register to the first location at which
2616 outgoing arguments are placed. If not specified, the default value of
2617 zero is used. This is the proper value for most machines.
2619 If @code{ARGS_GROW_DOWNWARD}, this is the offset to the location above
2620 the first location at which outgoing arguments are placed.
2622 @findex FIRST_PARM_OFFSET
2623 @item FIRST_PARM_OFFSET (@var{fundecl})
2624 Offset from the argument pointer register to the first argument's
2625 address. On some machines it may depend on the data type of the
2628 If @code{ARGS_GROW_DOWNWARD}, this is the offset to the location above
2629 the first argument's address.
2631 @findex STACK_DYNAMIC_OFFSET
2632 @item STACK_DYNAMIC_OFFSET (@var{fundecl})
2633 Offset from the stack pointer register to an item dynamically allocated
2634 on the stack, e.g., by @code{alloca}.
2636 The default value for this macro is @code{STACK_POINTER_OFFSET} plus the
2637 length of the outgoing arguments. The default is correct for most
2638 machines. See @file{function.c} for details.
2640 @findex DYNAMIC_CHAIN_ADDRESS
2641 @item DYNAMIC_CHAIN_ADDRESS (@var{frameaddr})
2642 A C expression whose value is RTL representing the address in a stack
2643 frame where the pointer to the caller's frame is stored. Assume that
2644 @var{frameaddr} is an RTL expression for the address of the stack frame
2647 If you don't define this macro, the default is to return the value
2648 of @var{frameaddr}---that is, the stack frame address is also the
2649 address of the stack word that points to the previous frame.
2651 @findex SETUP_FRAME_ADDRESSES
2652 @item SETUP_FRAME_ADDRESSES
2653 If defined, a C expression that produces the machine-specific code to
2654 setup the stack so that arbitrary frames can be accessed. For example,
2655 on the Sparc, we must flush all of the register windows to the stack
2656 before we can access arbitrary stack frames. You will seldom need to
2659 @findex BUILTIN_SETJMP_FRAME_VALUE
2660 @item BUILTIN_SETJMP_FRAME_VALUE
2661 If defined, a C expression that contains an rtx that is used to store
2662 the address of the current frame into the built in @code{setjmp} buffer.
2663 The default value, @code{virtual_stack_vars_rtx}, is correct for most
2664 machines. One reason you may need to define this macro is if
2665 @code{hard_frame_pointer_rtx} is the appropriate value on your machine.
2667 @findex RETURN_ADDR_RTX
2668 @item RETURN_ADDR_RTX (@var{count}, @var{frameaddr})
2669 A C expression whose value is RTL representing the value of the return
2670 address for the frame @var{count} steps up from the current frame, after
2671 the prologue. @var{frameaddr} is the frame pointer of the @var{count}
2672 frame, or the frame pointer of the @var{count} @minus{} 1 frame if
2673 @code{RETURN_ADDR_IN_PREVIOUS_FRAME} is defined.
2675 The value of the expression must always be the correct address when
2676 @var{count} is zero, but may be @code{NULL_RTX} if there is not way to
2677 determine the return address of other frames.
2679 @findex RETURN_ADDR_IN_PREVIOUS_FRAME
2680 @item RETURN_ADDR_IN_PREVIOUS_FRAME
2681 Define this if the return address of a particular stack frame is accessed
2682 from the frame pointer of the previous stack frame.
2684 @findex INCOMING_RETURN_ADDR_RTX
2685 @item INCOMING_RETURN_ADDR_RTX
2686 A C expression whose value is RTL representing the location of the
2687 incoming return address at the beginning of any function, before the
2688 prologue. This RTL is either a @code{REG}, indicating that the return
2689 value is saved in @samp{REG}, or a @code{MEM} representing a location in
2692 You only need to define this macro if you want to support call frame
2693 debugging information like that provided by DWARF 2.
2695 If this RTL is a @code{REG}, you should also define
2696 @code{DWARF_FRAME_RETURN_COLUMN} to @code{DWARF_FRAME_REGNUM (REGNO)}.
2698 @findex INCOMING_FRAME_SP_OFFSET
2699 @item INCOMING_FRAME_SP_OFFSET
2700 A C expression whose value is an integer giving the offset, in bytes,
2701 from the value of the stack pointer register to the top of the stack
2702 frame at the beginning of any function, before the prologue. The top of
2703 the frame is defined to be the value of the stack pointer in the
2704 previous frame, just before the call instruction.
2706 You only need to define this macro if you want to support call frame
2707 debugging information like that provided by DWARF 2.
2709 @findex ARG_POINTER_CFA_OFFSET
2710 @item ARG_POINTER_CFA_OFFSET (@var{fundecl})
2711 A C expression whose value is an integer giving the offset, in bytes,
2712 from the argument pointer to the canonical frame address (cfa). The
2713 final value should coincide with that calculated by
2714 @code{INCOMING_FRAME_SP_OFFSET}. Which is unfortunately not usable
2715 during virtual register instantiation.
2717 The default value for this macro is @code{FIRST_PARM_OFFSET (fundecl)},
2718 which is correct for most machines; in general, the arguments are found
2719 immediately before the stack frame. Note that this is not the case on
2720 some targets that save registers into the caller's frame, such as SPARC
2721 and rs6000, and so such targets need to define this macro.
2723 You only need to define this macro if the default is incorrect, and you
2724 want to support call frame debugging information like that provided by
2729 Define this macro if the stack size for the target is very small. This
2730 has the effect of disabling gcc's built-in @samp{alloca}, though
2731 @samp{__builtin_alloca} is not affected.
2734 @node Exception Handling
2735 @subsection Exception Handling Support
2736 @cindex exception handling
2739 @findex EH_RETURN_DATA_REGNO
2740 @item EH_RETURN_DATA_REGNO (@var{N})
2741 A C expression whose value is the @var{N}th register number used for
2742 data by exception handlers, or @code{INVALID_REGNUM} if fewer than
2743 @var{N} registers are usable.
2745 The exception handling library routines communicate with the exception
2746 handlers via a set of agreed upon registers. Ideally these registers
2747 should be call-clobbered; it is possible to use call-saved registers,
2748 but may negatively impact code size. The target must support at least
2749 2 data registers, but should define 4 if there are enough free registers.
2751 You must define this macro if you want to support call frame exception
2752 handling like that provided by DWARF 2.
2754 @findex EH_RETURN_STACKADJ_RTX
2755 @item EH_RETURN_STACKADJ_RTX
2756 A C expression whose value is RTL representing a location in which
2757 to store a stack adjustment to be applied before function return.
2758 This is used to unwind the stack to an exception handler's call frame.
2759 It will be assigned zero on code paths that return normally.
2761 Typically this is a call-clobbered hard register that is otherwise
2762 untouched by the epilogue, but could also be a stack slot.
2764 You must define this macro if you want to support call frame exception
2765 handling like that provided by DWARF 2.
2767 @findex EH_RETURN_HANDLER_RTX
2768 @item EH_RETURN_HANDLER_RTX
2769 A C expression whose value is RTL representing a location in which
2770 to store the address of an exception handler to which we should
2771 return. It will not be assigned on code paths that return normally.
2773 Typically this is the location in the call frame at which the normal
2774 return address is stored. For targets that return by popping an
2775 address off the stack, this might be a memory address just below
2776 the @emph{target} call frame rather than inside the current call
2777 frame. @code{EH_RETURN_STACKADJ_RTX} will have already been assigned,
2778 so it may be used to calculate the location of the target call frame.
2780 Some targets have more complex requirements than storing to an
2781 address calculable during initial code generation. In that case
2782 the @code{eh_return} instruction pattern should be used instead.
2784 If you want to support call frame exception handling, you must
2785 define either this macro or the @code{eh_return} instruction pattern.
2787 @findex ASM_PREFERRED_EH_DATA_FORMAT
2788 @item ASM_PREFERRED_EH_DATA_FORMAT(@var{code}, @var{global})
2789 This macro chooses the encoding of pointers embedded in the exception
2790 handling sections. If at all possible, this should be defined such
2791 that the exception handling section will not require dynamic relocations,
2792 and so may be read-only.
2794 @var{code} is 0 for data, 1 for code labels, 2 for function pointers.
2795 @var{global} is true if the symbol may be affected by dynamic relocations.
2796 The macro should return a combination of the @code{DW_EH_PE_*} defines
2797 as found in @file{dwarf2.h}.
2799 If this macro is not defined, pointers will not be encoded but
2800 represented directly.
2802 @findex ASM_MAYBE_OUTPUT_ENCODED_ADDR_RTX
2803 @item ASM_MAYBE_OUTPUT_ENCODED_ADDR_RTX(@var{file}, @var{encoding}, @var{size}, @var{addr}, @var{done})
2804 This macro allows the target to emit whatever special magic is required
2805 to represent the encoding chosen by @code{ASM_PREFERRED_EH_DATA_FORMAT}.
2806 Generic code takes care of pc-relative and indirect encodings; this must
2807 be defined if the target uses text-relative or data-relative encodings.
2809 This is a C statement that branches to @var{done} if the format was
2810 handled. @var{encoding} is the format chosen, @var{size} is the number
2811 of bytes that the format occupies, @var{addr} is the @code{SYMBOL_REF}
2814 @findex MD_FALLBACK_FRAME_STATE_FOR
2815 @item MD_FALLBACK_FRAME_STATE_FOR(@var{context}, @var{fs}, @var{success})
2816 This macro allows the target to add cpu and operating system specific
2817 code to the call-frame unwinder for use when there is no unwind data
2818 available. The most common reason to implement this macro is to unwind
2819 through signal frames.
2821 This macro is called from @code{uw_frame_state_for} in @file{unwind-dw2.c}
2822 and @file{unwind-ia64.c}. @var{context} is an @code{_Unwind_Context};
2823 @var{fs} is an @code{_Unwind_FrameState}. Examine @code{context->ra}
2824 for the address of the code being executed and @code{context->cfa} for
2825 the stack pointer value. If the frame can be decoded, the register save
2826 addresses should be updated in @var{fs} and the macro should branch to
2827 @var{success}. If the frame cannot be decoded, the macro should do
2831 @node Stack Checking
2832 @subsection Specifying How Stack Checking is Done
2834 GCC will check that stack references are within the boundaries of
2835 the stack, if the @option{-fstack-check} is specified, in one of three ways:
2839 If the value of the @code{STACK_CHECK_BUILTIN} macro is nonzero, GCC
2840 will assume that you have arranged for stack checking to be done at
2841 appropriate places in the configuration files, e.g., in
2842 @code{TARGET_ASM_FUNCTION_PROLOGUE}. GCC will do not other special
2846 If @code{STACK_CHECK_BUILTIN} is zero and you defined a named pattern
2847 called @code{check_stack} in your @file{md} file, GCC will call that
2848 pattern with one argument which is the address to compare the stack
2849 value against. You must arrange for this pattern to report an error if
2850 the stack pointer is out of range.
2853 If neither of the above are true, GCC will generate code to periodically
2854 ``probe'' the stack pointer using the values of the macros defined below.
2857 Normally, you will use the default values of these macros, so GCC
2858 will use the third approach.
2861 @findex STACK_CHECK_BUILTIN
2862 @item STACK_CHECK_BUILTIN
2863 A nonzero value if stack checking is done by the configuration files in a
2864 machine-dependent manner. You should define this macro if stack checking
2865 is require by the ABI of your machine or if you would like to have to stack
2866 checking in some more efficient way than GCC's portable approach.
2867 The default value of this macro is zero.
2869 @findex STACK_CHECK_PROBE_INTERVAL
2870 @item STACK_CHECK_PROBE_INTERVAL
2871 An integer representing the interval at which GCC must generate stack
2872 probe instructions. You will normally define this macro to be no larger
2873 than the size of the ``guard pages'' at the end of a stack area. The
2874 default value of 4096 is suitable for most systems.
2876 @findex STACK_CHECK_PROBE_LOAD
2877 @item STACK_CHECK_PROBE_LOAD
2878 A integer which is nonzero if GCC should perform the stack probe
2879 as a load instruction and zero if GCC should use a store instruction.
2880 The default is zero, which is the most efficient choice on most systems.
2882 @findex STACK_CHECK_PROTECT
2883 @item STACK_CHECK_PROTECT
2884 The number of bytes of stack needed to recover from a stack overflow,
2885 for languages where such a recovery is supported. The default value of
2886 75 words should be adequate for most machines.
2888 @findex STACK_CHECK_MAX_FRAME_SIZE
2889 @item STACK_CHECK_MAX_FRAME_SIZE
2890 The maximum size of a stack frame, in bytes. GCC will generate probe
2891 instructions in non-leaf functions to ensure at least this many bytes of
2892 stack are available. If a stack frame is larger than this size, stack
2893 checking will not be reliable and GCC will issue a warning. The
2894 default is chosen so that GCC only generates one instruction on most
2895 systems. You should normally not change the default value of this macro.
2897 @findex STACK_CHECK_FIXED_FRAME_SIZE
2898 @item STACK_CHECK_FIXED_FRAME_SIZE
2899 GCC uses this value to generate the above warning message. It
2900 represents the amount of fixed frame used by a function, not including
2901 space for any callee-saved registers, temporaries and user variables.
2902 You need only specify an upper bound for this amount and will normally
2903 use the default of four words.
2905 @findex STACK_CHECK_MAX_VAR_SIZE
2906 @item STACK_CHECK_MAX_VAR_SIZE
2907 The maximum size, in bytes, of an object that GCC will place in the
2908 fixed area of the stack frame when the user specifies
2909 @option{-fstack-check}.
2910 GCC computed the default from the values of the above macros and you will
2911 normally not need to override that default.
2915 @node Frame Registers
2916 @subsection Registers That Address the Stack Frame
2918 @c prevent bad page break with this line
2919 This discusses registers that address the stack frame.
2922 @findex STACK_POINTER_REGNUM
2923 @item STACK_POINTER_REGNUM
2924 The register number of the stack pointer register, which must also be a
2925 fixed register according to @code{FIXED_REGISTERS}. On most machines,
2926 the hardware determines which register this is.
2928 @findex FRAME_POINTER_REGNUM
2929 @item FRAME_POINTER_REGNUM
2930 The register number of the frame pointer register, which is used to
2931 access automatic variables in the stack frame. On some machines, the
2932 hardware determines which register this is. On other machines, you can
2933 choose any register you wish for this purpose.
2935 @findex HARD_FRAME_POINTER_REGNUM
2936 @item HARD_FRAME_POINTER_REGNUM
2937 On some machines the offset between the frame pointer and starting
2938 offset of the automatic variables is not known until after register
2939 allocation has been done (for example, because the saved registers are
2940 between these two locations). On those machines, define
2941 @code{FRAME_POINTER_REGNUM} the number of a special, fixed register to
2942 be used internally until the offset is known, and define
2943 @code{HARD_FRAME_POINTER_REGNUM} to be the actual hard register number
2944 used for the frame pointer.
2946 You should define this macro only in the very rare circumstances when it
2947 is not possible to calculate the offset between the frame pointer and
2948 the automatic variables until after register allocation has been
2949 completed. When this macro is defined, you must also indicate in your
2950 definition of @code{ELIMINABLE_REGS} how to eliminate
2951 @code{FRAME_POINTER_REGNUM} into either @code{HARD_FRAME_POINTER_REGNUM}
2952 or @code{STACK_POINTER_REGNUM}.
2954 Do not define this macro if it would be the same as
2955 @code{FRAME_POINTER_REGNUM}.
2957 @findex ARG_POINTER_REGNUM
2958 @item ARG_POINTER_REGNUM
2959 The register number of the arg pointer register, which is used to access
2960 the function's argument list. On some machines, this is the same as the
2961 frame pointer register. On some machines, the hardware determines which
2962 register this is. On other machines, you can choose any register you
2963 wish for this purpose. If this is not the same register as the frame
2964 pointer register, then you must mark it as a fixed register according to
2965 @code{FIXED_REGISTERS}, or arrange to be able to eliminate it
2966 (@pxref{Elimination}).
2968 @findex RETURN_ADDRESS_POINTER_REGNUM
2969 @item RETURN_ADDRESS_POINTER_REGNUM
2970 The register number of the return address pointer register, which is used to
2971 access the current function's return address from the stack. On some
2972 machines, the return address is not at a fixed offset from the frame
2973 pointer or stack pointer or argument pointer. This register can be defined
2974 to point to the return address on the stack, and then be converted by
2975 @code{ELIMINABLE_REGS} into either the frame pointer or stack pointer.
2977 Do not define this macro unless there is no other way to get the return
2978 address from the stack.
2980 @findex STATIC_CHAIN_REGNUM
2981 @findex STATIC_CHAIN_INCOMING_REGNUM
2982 @item STATIC_CHAIN_REGNUM
2983 @itemx STATIC_CHAIN_INCOMING_REGNUM
2984 Register numbers used for passing a function's static chain pointer. If
2985 register windows are used, the register number as seen by the called
2986 function is @code{STATIC_CHAIN_INCOMING_REGNUM}, while the register
2987 number as seen by the calling function is @code{STATIC_CHAIN_REGNUM}. If
2988 these registers are the same, @code{STATIC_CHAIN_INCOMING_REGNUM} need
2991 The static chain register need not be a fixed register.
2993 If the static chain is passed in memory, these macros should not be
2994 defined; instead, the next two macros should be defined.
2996 @findex STATIC_CHAIN
2997 @findex STATIC_CHAIN_INCOMING
2999 @itemx STATIC_CHAIN_INCOMING
3000 If the static chain is passed in memory, these macros provide rtx giving
3001 @code{mem} expressions that denote where they are stored.
3002 @code{STATIC_CHAIN} and @code{STATIC_CHAIN_INCOMING} give the locations
3003 as seen by the calling and called functions, respectively. Often the former
3004 will be at an offset from the stack pointer and the latter at an offset from
3007 @findex stack_pointer_rtx
3008 @findex frame_pointer_rtx
3009 @findex arg_pointer_rtx
3010 The variables @code{stack_pointer_rtx}, @code{frame_pointer_rtx}, and
3011 @code{arg_pointer_rtx} will have been initialized prior to the use of these
3012 macros and should be used to refer to those items.
3014 If the static chain is passed in a register, the two previous macros should
3017 @findex DWARF_FRAME_REGISTERS
3018 @item DWARF_FRAME_REGISTERS
3019 This macro specifies the maximum number of hard registers that can be
3020 saved in a call frame. This is used to size data structures used in
3021 DWARF2 exception handling.
3023 Prior to GCC 3.0, this macro was needed in order to establish a stable
3024 exception handling ABI in the face of adding new hard registers for ISA
3025 extensions. In GCC 3.0 and later, the EH ABI is insulated from changes
3026 in the number of hard registers. Nevertheless, this macro can still be
3027 used to reduce the runtime memory requirements of the exception handling
3028 routines, which can be substantial if the ISA contains a lot of
3029 registers that are not call-saved.
3031 If this macro is not defined, it defaults to
3032 @code{FIRST_PSEUDO_REGISTER}.
3034 @findex PRE_GCC3_DWARF_FRAME_REGISTERS
3035 @item PRE_GCC3_DWARF_FRAME_REGISTERS
3037 This macro is similar to @code{DWARF_FRAME_REGISTERS}, but is provided
3038 for backward compatibility in pre GCC 3.0 compiled code.
3040 If this macro is not defined, it defaults to
3041 @code{DWARF_FRAME_REGISTERS}.
3046 @subsection Eliminating Frame Pointer and Arg Pointer
3048 @c prevent bad page break with this line
3049 This is about eliminating the frame pointer and arg pointer.
3052 @findex FRAME_POINTER_REQUIRED
3053 @item FRAME_POINTER_REQUIRED
3054 A C expression which is nonzero if a function must have and use a frame
3055 pointer. This expression is evaluated in the reload pass. If its value is
3056 nonzero the function will have a frame pointer.
3058 The expression can in principle examine the current function and decide
3059 according to the facts, but on most machines the constant 0 or the
3060 constant 1 suffices. Use 0 when the machine allows code to be generated
3061 with no frame pointer, and doing so saves some time or space. Use 1
3062 when there is no possible advantage to avoiding a frame pointer.
3064 In certain cases, the compiler does not know how to produce valid code
3065 without a frame pointer. The compiler recognizes those cases and
3066 automatically gives the function a frame pointer regardless of what
3067 @code{FRAME_POINTER_REQUIRED} says. You don't need to worry about
3070 In a function that does not require a frame pointer, the frame pointer
3071 register can be allocated for ordinary usage, unless you mark it as a
3072 fixed register. See @code{FIXED_REGISTERS} for more information.
3074 @findex INITIAL_FRAME_POINTER_OFFSET
3075 @findex get_frame_size
3076 @item INITIAL_FRAME_POINTER_OFFSET (@var{depth-var})
3077 A C statement to store in the variable @var{depth-var} the difference
3078 between the frame pointer and the stack pointer values immediately after
3079 the function prologue. The value would be computed from information
3080 such as the result of @code{get_frame_size ()} and the tables of
3081 registers @code{regs_ever_live} and @code{call_used_regs}.
3083 If @code{ELIMINABLE_REGS} is defined, this macro will be not be used and
3084 need not be defined. Otherwise, it must be defined even if
3085 @code{FRAME_POINTER_REQUIRED} is defined to always be true; in that
3086 case, you may set @var{depth-var} to anything.
3088 @findex ELIMINABLE_REGS
3089 @item ELIMINABLE_REGS
3090 If defined, this macro specifies a table of register pairs used to
3091 eliminate unneeded registers that point into the stack frame. If it is not
3092 defined, the only elimination attempted by the compiler is to replace
3093 references to the frame pointer with references to the stack pointer.
3095 The definition of this macro is a list of structure initializations, each
3096 of which specifies an original and replacement register.
3098 On some machines, the position of the argument pointer is not known until
3099 the compilation is completed. In such a case, a separate hard register
3100 must be used for the argument pointer. This register can be eliminated by
3101 replacing it with either the frame pointer or the argument pointer,
3102 depending on whether or not the frame pointer has been eliminated.
3104 In this case, you might specify:
3106 #define ELIMINABLE_REGS \
3107 @{@{ARG_POINTER_REGNUM, STACK_POINTER_REGNUM@}, \
3108 @{ARG_POINTER_REGNUM, FRAME_POINTER_REGNUM@}, \
3109 @{FRAME_POINTER_REGNUM, STACK_POINTER_REGNUM@}@}
3112 Note that the elimination of the argument pointer with the stack pointer is
3113 specified first since that is the preferred elimination.
3115 @findex CAN_ELIMINATE
3116 @item CAN_ELIMINATE (@var{from-reg}, @var{to-reg})
3117 A C expression that returns nonzero if the compiler is allowed to try
3118 to replace register number @var{from-reg} with register number
3119 @var{to-reg}. This macro need only be defined if @code{ELIMINABLE_REGS}
3120 is defined, and will usually be the constant 1, since most of the cases
3121 preventing register elimination are things that the compiler already
3124 @findex INITIAL_ELIMINATION_OFFSET
3125 @item INITIAL_ELIMINATION_OFFSET (@var{from-reg}, @var{to-reg}, @var{offset-var})
3126 This macro is similar to @code{INITIAL_FRAME_POINTER_OFFSET}. It
3127 specifies the initial difference between the specified pair of
3128 registers. This macro must be defined if @code{ELIMINABLE_REGS} is
3132 @node Stack Arguments
3133 @subsection Passing Function Arguments on the Stack
3134 @cindex arguments on stack
3135 @cindex stack arguments
3137 The macros in this section control how arguments are passed
3138 on the stack. See the following section for other macros that
3139 control passing certain arguments in registers.
3142 @findex PROMOTE_PROTOTYPES
3143 @item PROMOTE_PROTOTYPES
3144 A C expression whose value is nonzero if an argument declared in
3145 a prototype as an integral type smaller than @code{int} should
3146 actually be passed as an @code{int}. In addition to avoiding
3147 errors in certain cases of mismatch, it also makes for better
3148 code on certain machines. If the macro is not defined in target
3149 header files, it defaults to 0.
3153 A C expression. If nonzero, push insns will be used to pass
3155 If the target machine does not have a push instruction, set it to zero.
3156 That directs GCC to use an alternate strategy: to
3157 allocate the entire argument block and then store the arguments into
3158 it. When @code{PUSH_ARGS} is nonzero, @code{PUSH_ROUNDING} must be defined too.
3159 On some machines, the definition
3161 @findex PUSH_ROUNDING
3162 @item PUSH_ROUNDING (@var{npushed})
3163 A C expression that is the number of bytes actually pushed onto the
3164 stack when an instruction attempts to push @var{npushed} bytes.
3166 On some machines, the definition
3169 #define PUSH_ROUNDING(BYTES) (BYTES)
3173 will suffice. But on other machines, instructions that appear
3174 to push one byte actually push two bytes in an attempt to maintain
3175 alignment. Then the definition should be
3178 #define PUSH_ROUNDING(BYTES) (((BYTES) + 1) & ~1)
3181 @findex ACCUMULATE_OUTGOING_ARGS
3182 @findex current_function_outgoing_args_size
3183 @item ACCUMULATE_OUTGOING_ARGS
3184 A C expression. If nonzero, the maximum amount of space required for outgoing arguments
3185 will be computed and placed into the variable
3186 @code{current_function_outgoing_args_size}. No space will be pushed
3187 onto the stack for each call; instead, the function prologue should
3188 increase the stack frame size by this amount.
3190 Setting both @code{PUSH_ARGS} and @code{ACCUMULATE_OUTGOING_ARGS}
3193 @findex REG_PARM_STACK_SPACE
3194 @item REG_PARM_STACK_SPACE (@var{fndecl})
3195 Define this macro if functions should assume that stack space has been
3196 allocated for arguments even when their values are passed in
3199 The value of this macro is the size, in bytes, of the area reserved for
3200 arguments passed in registers for the function represented by @var{fndecl},
3201 which can be zero if GCC is calling a library function.
3203 This space can be allocated by the caller, or be a part of the
3204 machine-dependent stack frame: @code{OUTGOING_REG_PARM_STACK_SPACE} says
3206 @c above is overfull. not sure what to do. --mew 5feb93 did
3207 @c something, not sure if it looks good. --mew 10feb93
3209 @findex MAYBE_REG_PARM_STACK_SPACE
3210 @findex FINAL_REG_PARM_STACK_SPACE
3211 @item MAYBE_REG_PARM_STACK_SPACE
3212 @itemx FINAL_REG_PARM_STACK_SPACE (@var{const_size}, @var{var_size})
3213 Define these macros in addition to the one above if functions might
3214 allocate stack space for arguments even when their values are passed
3215 in registers. These should be used when the stack space allocated
3216 for arguments in registers is not a simple constant independent of the
3217 function declaration.
3219 The value of the first macro is the size, in bytes, of the area that
3220 we should initially assume would be reserved for arguments passed in registers.
3222 The value of the second macro is the actual size, in bytes, of the area
3223 that will be reserved for arguments passed in registers. This takes two
3224 arguments: an integer representing the number of bytes of fixed sized
3225 arguments on the stack, and a tree representing the number of bytes of
3226 variable sized arguments on the stack.
3228 When these macros are defined, @code{REG_PARM_STACK_SPACE} will only be
3229 called for libcall functions, the current function, or for a function
3230 being called when it is known that such stack space must be allocated.
3231 In each case this value can be easily computed.
3233 When deciding whether a called function needs such stack space, and how
3234 much space to reserve, GCC uses these two macros instead of
3235 @code{REG_PARM_STACK_SPACE}.
3237 @findex OUTGOING_REG_PARM_STACK_SPACE
3238 @item OUTGOING_REG_PARM_STACK_SPACE
3239 Define this if it is the responsibility of the caller to allocate the area
3240 reserved for arguments passed in registers.
3242 If @code{ACCUMULATE_OUTGOING_ARGS} is defined, this macro controls
3243 whether the space for these arguments counts in the value of
3244 @code{current_function_outgoing_args_size}.
3246 @findex STACK_PARMS_IN_REG_PARM_AREA
3247 @item STACK_PARMS_IN_REG_PARM_AREA
3248 Define this macro if @code{REG_PARM_STACK_SPACE} is defined, but the
3249 stack parameters don't skip the area specified by it.
3250 @c i changed this, makes more sens and it should have taken care of the
3251 @c overfull.. not as specific, tho. --mew 5feb93
3253 Normally, when a parameter is not passed in registers, it is placed on the
3254 stack beyond the @code{REG_PARM_STACK_SPACE} area. Defining this macro
3255 suppresses this behavior and causes the parameter to be passed on the
3256 stack in its natural location.
3258 @findex RETURN_POPS_ARGS
3259 @item RETURN_POPS_ARGS (@var{fundecl}, @var{funtype}, @var{stack-size})
3260 A C expression that should indicate the number of bytes of its own
3261 arguments that a function pops on returning, or 0 if the
3262 function pops no arguments and the caller must therefore pop them all
3263 after the function returns.
3265 @var{fundecl} is a C variable whose value is a tree node that describes
3266 the function in question. Normally it is a node of type
3267 @code{FUNCTION_DECL} that describes the declaration of the function.
3268 From this you can obtain the @code{DECL_ATTRIBUTES} of the function.
3270 @var{funtype} is a C variable whose value is a tree node that
3271 describes the function in question. Normally it is a node of type
3272 @code{FUNCTION_TYPE} that describes the data type of the function.
3273 From this it is possible to obtain the data types of the value and
3274 arguments (if known).
3276 When a call to a library function is being considered, @var{fundecl}
3277 will contain an identifier node for the library function. Thus, if
3278 you need to distinguish among various library functions, you can do so
3279 by their names. Note that ``library function'' in this context means
3280 a function used to perform arithmetic, whose name is known specially
3281 in the compiler and was not mentioned in the C code being compiled.
3283 @var{stack-size} is the number of bytes of arguments passed on the
3284 stack. If a variable number of bytes is passed, it is zero, and
3285 argument popping will always be the responsibility of the calling function.
3287 On the VAX, all functions always pop their arguments, so the definition
3288 of this macro is @var{stack-size}. On the 68000, using the standard
3289 calling convention, no functions pop their arguments, so the value of
3290 the macro is always 0 in this case. But an alternative calling
3291 convention is available in which functions that take a fixed number of
3292 arguments pop them but other functions (such as @code{printf}) pop
3293 nothing (the caller pops all). When this convention is in use,
3294 @var{funtype} is examined to determine whether a function takes a fixed
3295 number of arguments.
3298 @node Register Arguments
3299 @subsection Passing Arguments in Registers
3300 @cindex arguments in registers
3301 @cindex registers arguments
3303 This section describes the macros which let you control how various
3304 types of arguments are passed in registers or how they are arranged in
3308 @findex FUNCTION_ARG
3309 @item FUNCTION_ARG (@var{cum}, @var{mode}, @var{type}, @var{named})
3310 A C expression that controls whether a function argument is passed
3311 in a register, and which register.
3313 The arguments are @var{cum}, which summarizes all the previous
3314 arguments; @var{mode}, the machine mode of the argument; @var{type},
3315 the data type of the argument as a tree node or 0 if that is not known
3316 (which happens for C support library functions); and @var{named},
3317 which is 1 for an ordinary argument and 0 for nameless arguments that
3318 correspond to @samp{@dots{}} in the called function's prototype.
3319 @var{type} can be an incomplete type if a syntax error has previously
3322 The value of the expression is usually either a @code{reg} RTX for the
3323 hard register in which to pass the argument, or zero to pass the
3324 argument on the stack.
3326 For machines like the VAX and 68000, where normally all arguments are
3327 pushed, zero suffices as a definition.
3329 The value of the expression can also be a @code{parallel} RTX@. This is
3330 used when an argument is passed in multiple locations. The mode of the
3331 of the @code{parallel} should be the mode of the entire argument. The
3332 @code{parallel} holds any number of @code{expr_list} pairs; each one
3333 describes where part of the argument is passed. In each
3334 @code{expr_list} the first operand must be a @code{reg} RTX for the hard
3335 register in which to pass this part of the argument, and the mode of the
3336 register RTX indicates how large this part of the argument is. The
3337 second operand of the @code{expr_list} is a @code{const_int} which gives
3338 the offset in bytes into the entire argument of where this part starts.
3339 As a special exception the first @code{expr_list} in the @code{parallel}
3340 RTX may have a first operand of zero. This indicates that the entire
3341 argument is also stored on the stack.
3343 The last time this macro is called, it is called with @code{MODE ==
3344 VOIDmode}, and its result is passed to the @code{call} or @code{call_value}
3345 pattern as operands 2 and 3 respectively.
3347 @cindex @file{stdarg.h} and register arguments
3348 The usual way to make the ISO library @file{stdarg.h} work on a machine
3349 where some arguments are usually passed in registers, is to cause
3350 nameless arguments to be passed on the stack instead. This is done
3351 by making @code{FUNCTION_ARG} return 0 whenever @var{named} is 0.
3353 @cindex @code{MUST_PASS_IN_STACK}, and @code{FUNCTION_ARG}
3354 @cindex @code{REG_PARM_STACK_SPACE}, and @code{FUNCTION_ARG}
3355 You may use the macro @code{MUST_PASS_IN_STACK (@var{mode}, @var{type})}
3356 in the definition of this macro to determine if this argument is of a
3357 type that must be passed in the stack. If @code{REG_PARM_STACK_SPACE}
3358 is not defined and @code{FUNCTION_ARG} returns nonzero for such an
3359 argument, the compiler will abort. If @code{REG_PARM_STACK_SPACE} is
3360 defined, the argument will be computed in the stack and then loaded into
3363 @findex MUST_PASS_IN_STACK
3364 @item MUST_PASS_IN_STACK (@var{mode}, @var{type})
3365 Define as a C expression that evaluates to nonzero if we do not know how
3366 to pass TYPE solely in registers. The file @file{expr.h} defines a
3367 definition that is usually appropriate, refer to @file{expr.h} for additional
3370 @findex FUNCTION_INCOMING_ARG
3371 @item FUNCTION_INCOMING_ARG (@var{cum}, @var{mode}, @var{type}, @var{named})
3372 Define this macro if the target machine has ``register windows'', so
3373 that the register in which a function sees an arguments is not
3374 necessarily the same as the one in which the caller passed the
3377 For such machines, @code{FUNCTION_ARG} computes the register in which
3378 the caller passes the value, and @code{FUNCTION_INCOMING_ARG} should
3379 be defined in a similar fashion to tell the function being called
3380 where the arguments will arrive.
3382 If @code{FUNCTION_INCOMING_ARG} is not defined, @code{FUNCTION_ARG}
3383 serves both purposes.
3385 @findex FUNCTION_ARG_PARTIAL_NREGS
3386 @item FUNCTION_ARG_PARTIAL_NREGS (@var{cum}, @var{mode}, @var{type}, @var{named})
3387 A C expression for the number of words, at the beginning of an
3388 argument, that must be put in registers. The value must be zero for
3389 arguments that are passed entirely in registers or that are entirely
3390 pushed on the stack.
3392 On some machines, certain arguments must be passed partially in
3393 registers and partially in memory. On these machines, typically the
3394 first @var{n} words of arguments are passed in registers, and the rest
3395 on the stack. If a multi-word argument (a @code{double} or a
3396 structure) crosses that boundary, its first few words must be passed
3397 in registers and the rest must be pushed. This macro tells the
3398 compiler when this occurs, and how many of the words should go in
3401 @code{FUNCTION_ARG} for these arguments should return the first
3402 register to be used by the caller for this argument; likewise
3403 @code{FUNCTION_INCOMING_ARG}, for the called function.
3405 @findex FUNCTION_ARG_PASS_BY_REFERENCE
3406 @item FUNCTION_ARG_PASS_BY_REFERENCE (@var{cum}, @var{mode}, @var{type}, @var{named})
3407 A C expression that indicates when an argument must be passed by reference.
3408 If nonzero for an argument, a copy of that argument is made in memory and a
3409 pointer to the argument is passed instead of the argument itself.
3410 The pointer is passed in whatever way is appropriate for passing a pointer
3413 On machines where @code{REG_PARM_STACK_SPACE} is not defined, a suitable
3414 definition of this macro might be
3416 #define FUNCTION_ARG_PASS_BY_REFERENCE\
3417 (CUM, MODE, TYPE, NAMED) \
3418 MUST_PASS_IN_STACK (MODE, TYPE)
3420 @c this is *still* too long. --mew 5feb93
3422 @findex FUNCTION_ARG_CALLEE_COPIES
3423 @item FUNCTION_ARG_CALLEE_COPIES (@var{cum}, @var{mode}, @var{type}, @var{named})
3424 If defined, a C expression that indicates when it is the called function's
3425 responsibility to make a copy of arguments passed by invisible reference.
3426 Normally, the caller makes a copy and passes the address of the copy to the
3427 routine being called. When @code{FUNCTION_ARG_CALLEE_COPIES} is defined and is
3428 nonzero, the caller does not make a copy. Instead, it passes a pointer to the
3429 ``live'' value. The called function must not modify this value. If it can be
3430 determined that the value won't be modified, it need not make a copy;
3431 otherwise a copy must be made.
3433 @findex FUNCTION_ARG_REG_LITTLE_ENDIAN
3434 @item FUNCTION_ARG_REG_LITTLE_ENDIAN
3435 If defined TRUE on a big-endian system then structure arguments passed
3436 (and returned) in registers are passed in a little-endian manner instead of
3437 the big-endian manner. On the HP-UX IA64 and PA64 platforms structures are
3438 aligned differently then integral values and setting this value to true will
3439 allow for the special handling of structure arguments and return values.
3441 @findex CUMULATIVE_ARGS
3442 @item CUMULATIVE_ARGS
3443 A C type for declaring a variable that is used as the first argument of
3444 @code{FUNCTION_ARG} and other related values. For some target machines,
3445 the type @code{int} suffices and can hold the number of bytes of
3448 There is no need to record in @code{CUMULATIVE_ARGS} anything about the
3449 arguments that have been passed on the stack. The compiler has other
3450 variables to keep track of that. For target machines on which all
3451 arguments are passed on the stack, there is no need to store anything in
3452 @code{CUMULATIVE_ARGS}; however, the data structure must exist and
3453 should not be empty, so use @code{int}.
3455 @findex INIT_CUMULATIVE_ARGS
3456 @item INIT_CUMULATIVE_ARGS (@var{cum}, @var{fntype}, @var{libname}, @var{indirect})
3457 A C statement (sans semicolon) for initializing the variable @var{cum}
3458 for the state at the beginning of the argument list. The variable has
3459 type @code{CUMULATIVE_ARGS}. The value of @var{fntype} is the tree node
3460 for the data type of the function which will receive the args, or 0
3461 if the args are to a compiler support library function. The value of
3462 @var{indirect} is nonzero when processing an indirect call, for example
3463 a call through a function pointer. The value of @var{indirect} is zero
3464 for a call to an explicitly named function, a library function call, or when
3465 @code{INIT_CUMULATIVE_ARGS} is used to find arguments for the function
3468 When processing a call to a compiler support library function,
3469 @var{libname} identifies which one. It is a @code{symbol_ref} rtx which
3470 contains the name of the function, as a string. @var{libname} is 0 when
3471 an ordinary C function call is being processed. Thus, each time this
3472 macro is called, either @var{libname} or @var{fntype} is nonzero, but
3473 never both of them at once.
3475 @findex INIT_CUMULATIVE_LIBCALL_ARGS
3476 @item INIT_CUMULATIVE_LIBCALL_ARGS (@var{cum}, @var{mode}, @var{libname})
3477 Like @code{INIT_CUMULATIVE_ARGS} but only used for outgoing libcalls,
3478 it gets a @code{MODE} argument instead of @var{fntype}, that would be
3479 @code{NULL}. @var{indirect} would always be zero, too. If this macro
3480 is not defined, @code{INIT_CUMULATIVE_ARGS (cum, NULL_RTX, libname,
3481 0)} is used instead.
3483 @findex INIT_CUMULATIVE_INCOMING_ARGS
3484 @item INIT_CUMULATIVE_INCOMING_ARGS (@var{cum}, @var{fntype}, @var{libname})
3485 Like @code{INIT_CUMULATIVE_ARGS} but overrides it for the purposes of
3486 finding the arguments for the function being compiled. If this macro is
3487 undefined, @code{INIT_CUMULATIVE_ARGS} is used instead.
3489 The value passed for @var{libname} is always 0, since library routines
3490 with special calling conventions are never compiled with GCC@. The
3491 argument @var{libname} exists for symmetry with
3492 @code{INIT_CUMULATIVE_ARGS}.
3493 @c could use "this macro" in place of @code{INIT_CUMULATIVE_ARGS}, maybe.
3494 @c --mew 5feb93 i switched the order of the sentences. --mew 10feb93
3496 @findex FUNCTION_ARG_ADVANCE
3497 @item FUNCTION_ARG_ADVANCE (@var{cum}, @var{mode}, @var{type}, @var{named})
3498 A C statement (sans semicolon) to update the summarizer variable
3499 @var{cum} to advance past an argument in the argument list. The
3500 values @var{mode}, @var{type} and @var{named} describe that argument.
3501 Once this is done, the variable @var{cum} is suitable for analyzing
3502 the @emph{following} argument with @code{FUNCTION_ARG}, etc.
3504 This macro need not do anything if the argument in question was passed
3505 on the stack. The compiler knows how to track the amount of stack space
3506 used for arguments without any special help.
3508 @findex FUNCTION_ARG_PADDING
3509 @item FUNCTION_ARG_PADDING (@var{mode}, @var{type})
3510 If defined, a C expression which determines whether, and in which direction,
3511 to pad out an argument with extra space. The value should be of type
3512 @code{enum direction}: either @code{upward} to pad above the argument,
3513 @code{downward} to pad below, or @code{none} to inhibit padding.
3515 The @emph{amount} of padding is always just enough to reach the next
3516 multiple of @code{FUNCTION_ARG_BOUNDARY}; this macro does not control
3519 This macro has a default definition which is right for most systems.
3520 For little-endian machines, the default is to pad upward. For
3521 big-endian machines, the default is to pad downward for an argument of
3522 constant size shorter than an @code{int}, and upward otherwise.
3524 @findex PAD_VARARGS_DOWN
3525 @item PAD_VARARGS_DOWN
3526 If defined, a C expression which determines whether the default
3527 implementation of va_arg will attempt to pad down before reading the
3528 next argument, if that argument is smaller than its aligned space as
3529 controlled by @code{PARM_BOUNDARY}. If this macro is not defined, all such
3530 arguments are padded down if @code{BYTES_BIG_ENDIAN} is true.
3532 @findex FUNCTION_ARG_BOUNDARY
3533 @item FUNCTION_ARG_BOUNDARY (@var{mode}, @var{type})
3534 If defined, a C expression that gives the alignment boundary, in bits,
3535 of an argument with the specified mode and type. If it is not defined,
3536 @code{PARM_BOUNDARY} is used for all arguments.
3538 @findex FUNCTION_ARG_REGNO_P
3539 @item FUNCTION_ARG_REGNO_P (@var{regno})
3540 A C expression that is nonzero if @var{regno} is the number of a hard
3541 register in which function arguments are sometimes passed. This does
3542 @emph{not} include implicit arguments such as the static chain and
3543 the structure-value address. On many machines, no registers can be
3544 used for this purpose since all function arguments are pushed on the
3547 @findex LOAD_ARGS_REVERSED
3548 @item LOAD_ARGS_REVERSED
3549 If defined, the order in which arguments are loaded into their
3550 respective argument registers is reversed so that the last
3551 argument is loaded first. This macro only affects arguments
3552 passed in registers.
3557 @subsection How Scalar Function Values Are Returned
3558 @cindex return values in registers
3559 @cindex values, returned by functions
3560 @cindex scalars, returned as values
3562 This section discusses the macros that control returning scalars as
3563 values---values that can fit in registers.
3566 @findex TRADITIONAL_RETURN_FLOAT
3567 @item TRADITIONAL_RETURN_FLOAT
3568 Define this macro if @option{-traditional} should not cause functions
3569 declared to return @code{float} to convert the value to @code{double}.
3571 @findex FUNCTION_VALUE
3572 @item FUNCTION_VALUE (@var{valtype}, @var{func})
3573 A C expression to create an RTX representing the place where a
3574 function returns a value of data type @var{valtype}. @var{valtype} is
3575 a tree node representing a data type. Write @code{TYPE_MODE
3576 (@var{valtype})} to get the machine mode used to represent that type.
3577 On many machines, only the mode is relevant. (Actually, on most
3578 machines, scalar values are returned in the same place regardless of
3581 The value of the expression is usually a @code{reg} RTX for the hard
3582 register where the return value is stored. The value can also be a
3583 @code{parallel} RTX, if the return value is in multiple places. See
3584 @code{FUNCTION_ARG} for an explanation of the @code{parallel} form.
3586 If @code{PROMOTE_FUNCTION_RETURN} is defined, you must apply the same
3587 promotion rules specified in @code{PROMOTE_MODE} if @var{valtype} is a
3590 If the precise function being called is known, @var{func} is a tree
3591 node (@code{FUNCTION_DECL}) for it; otherwise, @var{func} is a null
3592 pointer. This makes it possible to use a different value-returning
3593 convention for specific functions when all their calls are
3596 @code{FUNCTION_VALUE} is not used for return vales with aggregate data
3597 types, because these are returned in another way. See
3598 @code{STRUCT_VALUE_REGNUM} and related macros, below.
3600 @findex FUNCTION_OUTGOING_VALUE
3601 @item FUNCTION_OUTGOING_VALUE (@var{valtype}, @var{func})
3602 Define this macro if the target machine has ``register windows''
3603 so that the register in which a function returns its value is not
3604 the same as the one in which the caller sees the value.
3606 For such machines, @code{FUNCTION_VALUE} computes the register in which
3607 the caller will see the value. @code{FUNCTION_OUTGOING_VALUE} should be
3608 defined in a similar fashion to tell the function where to put the
3611 If @code{FUNCTION_OUTGOING_VALUE} is not defined,
3612 @code{FUNCTION_VALUE} serves both purposes.
3614 @code{FUNCTION_OUTGOING_VALUE} is not used for return vales with
3615 aggregate data types, because these are returned in another way. See
3616 @code{STRUCT_VALUE_REGNUM} and related macros, below.
3618 @findex LIBCALL_VALUE
3619 @item LIBCALL_VALUE (@var{mode})
3620 A C expression to create an RTX representing the place where a library
3621 function returns a value of mode @var{mode}. If the precise function
3622 being called is known, @var{func} is a tree node
3623 (@code{FUNCTION_DECL}) for it; otherwise, @var{func} is a null
3624 pointer. This makes it possible to use a different value-returning
3625 convention for specific functions when all their calls are
3628 Note that ``library function'' in this context means a compiler
3629 support routine, used to perform arithmetic, whose name is known
3630 specially by the compiler and was not mentioned in the C code being
3633 The definition of @code{LIBRARY_VALUE} need not be concerned aggregate
3634 data types, because none of the library functions returns such types.
3636 @findex FUNCTION_VALUE_REGNO_P
3637 @item FUNCTION_VALUE_REGNO_P (@var{regno})
3638 A C expression that is nonzero if @var{regno} is the number of a hard
3639 register in which the values of called function may come back.
3641 A register whose use for returning values is limited to serving as the
3642 second of a pair (for a value of type @code{double}, say) need not be
3643 recognized by this macro. So for most machines, this definition
3647 #define FUNCTION_VALUE_REGNO_P(N) ((N) == 0)
3650 If the machine has register windows, so that the caller and the called
3651 function use different registers for the return value, this macro
3652 should recognize only the caller's register numbers.
3654 @findex APPLY_RESULT_SIZE
3655 @item APPLY_RESULT_SIZE
3656 Define this macro if @samp{untyped_call} and @samp{untyped_return}
3657 need more space than is implied by @code{FUNCTION_VALUE_REGNO_P} for
3658 saving and restoring an arbitrary return value.
3661 @node Aggregate Return
3662 @subsection How Large Values Are Returned
3663 @cindex aggregates as return values
3664 @cindex large return values
3665 @cindex returning aggregate values
3666 @cindex structure value address
3668 When a function value's mode is @code{BLKmode} (and in some other
3669 cases), the value is not returned according to @code{FUNCTION_VALUE}
3670 (@pxref{Scalar Return}). Instead, the caller passes the address of a
3671 block of memory in which the value should be stored. This address
3672 is called the @dfn{structure value address}.
3674 This section describes how to control returning structure values in
3678 @findex RETURN_IN_MEMORY
3679 @item RETURN_IN_MEMORY (@var{type})
3680 A C expression which can inhibit the returning of certain function
3681 values in registers, based on the type of value. A nonzero value says
3682 to return the function value in memory, just as large structures are
3683 always returned. Here @var{type} will be a C expression of type
3684 @code{tree}, representing the data type of the value.
3686 Note that values of mode @code{BLKmode} must be explicitly handled
3687 by this macro. Also, the option @option{-fpcc-struct-return}
3688 takes effect regardless of this macro. On most systems, it is
3689 possible to leave the macro undefined; this causes a default
3690 definition to be used, whose value is the constant 1 for @code{BLKmode}
3691 values, and 0 otherwise.
3693 Do not use this macro to indicate that structures and unions should always
3694 be returned in memory. You should instead use @code{DEFAULT_PCC_STRUCT_RETURN}
3697 @findex DEFAULT_PCC_STRUCT_RETURN
3698 @item DEFAULT_PCC_STRUCT_RETURN
3699 Define this macro to be 1 if all structure and union return values must be
3700 in memory. Since this results in slower code, this should be defined
3701 only if needed for compatibility with other compilers or with an ABI@.
3702 If you define this macro to be 0, then the conventions used for structure
3703 and union return values are decided by the @code{RETURN_IN_MEMORY} macro.
3705 If not defined, this defaults to the value 1.
3707 @findex STRUCT_VALUE_REGNUM
3708 @item STRUCT_VALUE_REGNUM
3709 If the structure value address is passed in a register, then
3710 @code{STRUCT_VALUE_REGNUM} should be the number of that register.
3712 @findex STRUCT_VALUE
3714 If the structure value address is not passed in a register, define
3715 @code{STRUCT_VALUE} as an expression returning an RTX for the place
3716 where the address is passed. If it returns 0, the address is passed as
3717 an ``invisible'' first argument.
3719 @findex STRUCT_VALUE_INCOMING_REGNUM
3720 @item STRUCT_VALUE_INCOMING_REGNUM
3721 On some architectures the place where the structure value address
3722 is found by the called function is not the same place that the
3723 caller put it. This can be due to register windows, or it could
3724 be because the function prologue moves it to a different place.
3726 If the incoming location of the structure value address is in a
3727 register, define this macro as the register number.
3729 @findex STRUCT_VALUE_INCOMING
3730 @item STRUCT_VALUE_INCOMING
3731 If the incoming location is not a register, then you should define
3732 @code{STRUCT_VALUE_INCOMING} as an expression for an RTX for where the
3733 called function should find the value. If it should find the value on
3734 the stack, define this to create a @code{mem} which refers to the frame
3735 pointer. A definition of 0 means that the address is passed as an
3736 ``invisible'' first argument.
3738 @findex PCC_STATIC_STRUCT_RETURN
3739 @item PCC_STATIC_STRUCT_RETURN
3740 Define this macro if the usual system convention on the target machine
3741 for returning structures and unions is for the called function to return
3742 the address of a static variable containing the value.
3744 Do not define this if the usual system convention is for the caller to
3745 pass an address to the subroutine.
3747 This macro has effect in @option{-fpcc-struct-return} mode, but it does
3748 nothing when you use @option{-freg-struct-return} mode.
3752 @subsection Caller-Saves Register Allocation
3754 If you enable it, GCC can save registers around function calls. This
3755 makes it possible to use call-clobbered registers to hold variables that
3756 must live across calls.
3759 @findex DEFAULT_CALLER_SAVES
3760 @item DEFAULT_CALLER_SAVES
3761 Define this macro if function calls on the target machine do not preserve
3762 any registers; in other words, if @code{CALL_USED_REGISTERS} has 1
3763 for all registers. When defined, this macro enables @option{-fcaller-saves}
3764 by default for all optimization levels. It has no effect for optimization
3765 levels 2 and higher, where @option{-fcaller-saves} is the default.
3767 @findex CALLER_SAVE_PROFITABLE
3768 @item CALLER_SAVE_PROFITABLE (@var{refs}, @var{calls})
3769 A C expression to determine whether it is worthwhile to consider placing
3770 a pseudo-register in a call-clobbered hard register and saving and
3771 restoring it around each function call. The expression should be 1 when
3772 this is worth doing, and 0 otherwise.
3774 If you don't define this macro, a default is used which is good on most
3775 machines: @code{4 * @var{calls} < @var{refs}}.
3777 @findex HARD_REGNO_CALLER_SAVE_MODE
3778 @item HARD_REGNO_CALLER_SAVE_MODE (@var{regno}, @var{nregs})
3779 A C expression specifying which mode is required for saving @var{nregs}
3780 of a pseudo-register in call-clobbered hard register @var{regno}. If
3781 @var{regno} is unsuitable for caller save, @code{VOIDmode} should be
3782 returned. For most machines this macro need not be defined since GCC
3783 will select the smallest suitable mode.
3786 @node Function Entry
3787 @subsection Function Entry and Exit
3788 @cindex function entry and exit
3792 This section describes the macros that output function entry
3793 (@dfn{prologue}) and exit (@dfn{epilogue}) code.
3795 @deftypefn {Target Hook} void TARGET_ASM_FUNCTION_PROLOGUE (FILE *@var{file}, HOST_WIDE_INT @var{size})
3796 If defined, a function that outputs the assembler code for entry to a
3797 function. The prologue is responsible for setting up the stack frame,
3798 initializing the frame pointer register, saving registers that must be
3799 saved, and allocating @var{size} additional bytes of storage for the
3800 local variables. @var{size} is an integer. @var{file} is a stdio
3801 stream to which the assembler code should be output.
3803 The label for the beginning of the function need not be output by this
3804 macro. That has already been done when the macro is run.
3806 @findex regs_ever_live
3807 To determine which registers to save, the macro can refer to the array
3808 @code{regs_ever_live}: element @var{r} is nonzero if hard register
3809 @var{r} is used anywhere within the function. This implies the function
3810 prologue should save register @var{r}, provided it is not one of the
3811 call-used registers. (@code{TARGET_ASM_FUNCTION_EPILOGUE} must likewise use
3812 @code{regs_ever_live}.)
3814 On machines that have ``register windows'', the function entry code does
3815 not save on the stack the registers that are in the windows, even if
3816 they are supposed to be preserved by function calls; instead it takes
3817 appropriate steps to ``push'' the register stack, if any non-call-used
3818 registers are used in the function.
3820 @findex frame_pointer_needed
3821 On machines where functions may or may not have frame-pointers, the
3822 function entry code must vary accordingly; it must set up the frame
3823 pointer if one is wanted, and not otherwise. To determine whether a
3824 frame pointer is in wanted, the macro can refer to the variable
3825 @code{frame_pointer_needed}. The variable's value will be 1 at run
3826 time in a function that needs a frame pointer. @xref{Elimination}.
3828 The function entry code is responsible for allocating any stack space
3829 required for the function. This stack space consists of the regions
3830 listed below. In most cases, these regions are allocated in the
3831 order listed, with the last listed region closest to the top of the
3832 stack (the lowest address if @code{STACK_GROWS_DOWNWARD} is defined, and
3833 the highest address if it is not defined). You can use a different order
3834 for a machine if doing so is more convenient or required for
3835 compatibility reasons. Except in cases where required by standard
3836 or by a debugger, there is no reason why the stack layout used by GCC
3837 need agree with that used by other compilers for a machine.
3840 @deftypefn {Target Hook} void TARGET_ASM_FUNCTION_END_PROLOGUE (FILE *@var{file})
3841 If defined, a function that outputs assembler code at the end of a
3842 prologue. This should be used when the function prologue is being
3843 emitted as RTL, and you have some extra assembler that needs to be
3844 emitted. @xref{prologue instruction pattern}.
3847 @deftypefn {Target Hook} void TARGET_ASM_FUNCTION_BEGIN_EPILOGUE (FILE *@var{file})
3848 If defined, a function that outputs assembler code at the start of an
3849 epilogue. This should be used when the function epilogue is being
3850 emitted as RTL, and you have some extra assembler that needs to be
3851 emitted. @xref{epilogue instruction pattern}.
3854 @deftypefn {Target Hook} void TARGET_ASM_FUNCTION_EPILOGUE (FILE *@var{file}, HOST_WIDE_INT @var{size})
3855 If defined, a function that outputs the assembler code for exit from a
3856 function. The epilogue is responsible for restoring the saved
3857 registers and stack pointer to their values when the function was
3858 called, and returning control to the caller. This macro takes the
3859 same arguments as the macro @code{TARGET_ASM_FUNCTION_PROLOGUE}, and the
3860 registers to restore are determined from @code{regs_ever_live} and
3861 @code{CALL_USED_REGISTERS} in the same way.
3863 On some machines, there is a single instruction that does all the work
3864 of returning from the function. On these machines, give that
3865 instruction the name @samp{return} and do not define the macro
3866 @code{TARGET_ASM_FUNCTION_EPILOGUE} at all.
3868 Do not define a pattern named @samp{return} if you want the
3869 @code{TARGET_ASM_FUNCTION_EPILOGUE} to be used. If you want the target
3870 switches to control whether return instructions or epilogues are used,
3871 define a @samp{return} pattern with a validity condition that tests the
3872 target switches appropriately. If the @samp{return} pattern's validity
3873 condition is false, epilogues will be used.
3875 On machines where functions may or may not have frame-pointers, the
3876 function exit code must vary accordingly. Sometimes the code for these
3877 two cases is completely different. To determine whether a frame pointer
3878 is wanted, the macro can refer to the variable
3879 @code{frame_pointer_needed}. The variable's value will be 1 when compiling
3880 a function that needs a frame pointer.
3882 Normally, @code{TARGET_ASM_FUNCTION_PROLOGUE} and
3883 @code{TARGET_ASM_FUNCTION_EPILOGUE} must treat leaf functions specially.
3884 The C variable @code{current_function_is_leaf} is nonzero for such a
3885 function. @xref{Leaf Functions}.
3887 On some machines, some functions pop their arguments on exit while
3888 others leave that for the caller to do. For example, the 68020 when
3889 given @option{-mrtd} pops arguments in functions that take a fixed
3890 number of arguments.
3892 @findex current_function_pops_args
3893 Your definition of the macro @code{RETURN_POPS_ARGS} decides which
3894 functions pop their own arguments. @code{TARGET_ASM_FUNCTION_EPILOGUE}
3895 needs to know what was decided. The variable that is called
3896 @code{current_function_pops_args} is the number of bytes of its
3897 arguments that a function should pop. @xref{Scalar Return}.
3898 @c what is the "its arguments" in the above sentence referring to, pray
3899 @c tell? --mew 5feb93
3906 @findex current_function_pretend_args_size
3907 A region of @code{current_function_pretend_args_size} bytes of
3908 uninitialized space just underneath the first argument arriving on the
3909 stack. (This may not be at the very start of the allocated stack region
3910 if the calling sequence has pushed anything else since pushing the stack
3911 arguments. But usually, on such machines, nothing else has been pushed
3912 yet, because the function prologue itself does all the pushing.) This
3913 region is used on machines where an argument may be passed partly in
3914 registers and partly in memory, and, in some cases to support the
3915 features in @code{<varargs.h>} and @code{<stdarg.h>}.
3918 An area of memory used to save certain registers used by the function.
3919 The size of this area, which may also include space for such things as
3920 the return address and pointers to previous stack frames, is
3921 machine-specific and usually depends on which registers have been used
3922 in the function. Machines with register windows often do not require
3926 A region of at least @var{size} bytes, possibly rounded up to an allocation
3927 boundary, to contain the local variables of the function. On some machines,
3928 this region and the save area may occur in the opposite order, with the
3929 save area closer to the top of the stack.
3932 @cindex @code{ACCUMULATE_OUTGOING_ARGS} and stack frames
3933 Optionally, when @code{ACCUMULATE_OUTGOING_ARGS} is defined, a region of
3934 @code{current_function_outgoing_args_size} bytes to be used for outgoing
3935 argument lists of the function. @xref{Stack Arguments}.
3938 Normally, it is necessary for the macros
3939 @code{TARGET_ASM_FUNCTION_PROLOGUE} and
3940 @code{TARGET_ASM_FUNCTION_EPILOGUE} to treat leaf functions specially.
3941 The C variable @code{current_function_is_leaf} is nonzero for such a
3944 @findex EXIT_IGNORE_STACK
3945 @item EXIT_IGNORE_STACK
3946 Define this macro as a C expression that is nonzero if the return
3947 instruction or the function epilogue ignores the value of the stack
3948 pointer; in other words, if it is safe to delete an instruction to
3949 adjust the stack pointer before a return from the function.
3951 Note that this macro's value is relevant only for functions for which
3952 frame pointers are maintained. It is never safe to delete a final
3953 stack adjustment in a function that has no frame pointer, and the
3954 compiler knows this regardless of @code{EXIT_IGNORE_STACK}.
3956 @findex EPILOGUE_USES
3957 @item EPILOGUE_USES (@var{regno})
3958 Define this macro as a C expression that is nonzero for registers that are
3959 used by the epilogue or the @samp{return} pattern. The stack and frame
3960 pointer registers are already be assumed to be used as needed.
3962 @findex DELAY_SLOTS_FOR_EPILOGUE
3963 @item DELAY_SLOTS_FOR_EPILOGUE
3964 Define this macro if the function epilogue contains delay slots to which
3965 instructions from the rest of the function can be ``moved''. The
3966 definition should be a C expression whose value is an integer
3967 representing the number of delay slots there.
3969 @findex ELIGIBLE_FOR_EPILOGUE_DELAY
3970 @item ELIGIBLE_FOR_EPILOGUE_DELAY (@var{insn}, @var{n})
3971 A C expression that returns 1 if @var{insn} can be placed in delay
3972 slot number @var{n} of the epilogue.
3974 The argument @var{n} is an integer which identifies the delay slot now
3975 being considered (since different slots may have different rules of
3976 eligibility). It is never negative and is always less than the number
3977 of epilogue delay slots (what @code{DELAY_SLOTS_FOR_EPILOGUE} returns).
3978 If you reject a particular insn for a given delay slot, in principle, it
3979 may be reconsidered for a subsequent delay slot. Also, other insns may
3980 (at least in principle) be considered for the so far unfilled delay
3983 @findex current_function_epilogue_delay_list
3984 @findex final_scan_insn
3985 The insns accepted to fill the epilogue delay slots are put in an RTL
3986 list made with @code{insn_list} objects, stored in the variable
3987 @code{current_function_epilogue_delay_list}. The insn for the first
3988 delay slot comes first in the list. Your definition of the macro
3989 @code{TARGET_ASM_FUNCTION_EPILOGUE} should fill the delay slots by
3990 outputting the insns in this list, usually by calling
3991 @code{final_scan_insn}.
3993 You need not define this macro if you did not define
3994 @code{DELAY_SLOTS_FOR_EPILOGUE}.
3996 @findex ASM_OUTPUT_MI_THUNK
3997 @item ASM_OUTPUT_MI_THUNK (@var{file}, @var{thunk_fndecl}, @var{delta}, @var{function})
3998 A C compound statement that outputs the assembler code for a thunk
3999 function, used to implement C++ virtual function calls with multiple
4000 inheritance. The thunk acts as a wrapper around a virtual function,
4001 adjusting the implicit object parameter before handing control off to
4004 First, emit code to add the integer @var{delta} to the location that
4005 contains the incoming first argument. Assume that this argument
4006 contains a pointer, and is the one used to pass the @code{this} pointer
4007 in C++. This is the incoming argument @emph{before} the function prologue,
4008 e.g.@: @samp{%o0} on a sparc. The addition must preserve the values of
4009 all other incoming arguments.
4011 After the addition, emit code to jump to @var{function}, which is a
4012 @code{FUNCTION_DECL}. This is a direct pure jump, not a call, and does
4013 not touch the return address. Hence returning from @var{FUNCTION} will
4014 return to whoever called the current @samp{thunk}.
4016 The effect must be as if @var{function} had been called directly with
4017 the adjusted first argument. This macro is responsible for emitting all
4018 of the code for a thunk function; @code{TARGET_ASM_FUNCTION_PROLOGUE}
4019 and @code{TARGET_ASM_FUNCTION_EPILOGUE} are not invoked.
4021 The @var{thunk_fndecl} is redundant. (@var{delta} and @var{function}
4022 have already been extracted from it.) It might possibly be useful on
4023 some targets, but probably not.
4025 If you do not define this macro, the target-independent code in the C++
4026 front end will generate a less efficient heavyweight thunk that calls
4027 @var{function} instead of jumping to it. The generic approach does
4028 not support varargs.
4032 @subsection Generating Code for Profiling
4033 @cindex profiling, code generation
4035 These macros will help you generate code for profiling.
4038 @findex FUNCTION_PROFILER
4039 @item FUNCTION_PROFILER (@var{file}, @var{labelno})
4040 A C statement or compound statement to output to @var{file} some
4041 assembler code to call the profiling subroutine @code{mcount}.
4044 The details of how @code{mcount} expects to be called are determined by
4045 your operating system environment, not by GCC@. To figure them out,
4046 compile a small program for profiling using the system's installed C
4047 compiler and look at the assembler code that results.
4049 Older implementations of @code{mcount} expect the address of a counter
4050 variable to be loaded into some register. The name of this variable is
4051 @samp{LP} followed by the number @var{labelno}, so you would generate
4052 the name using @samp{LP%d} in a @code{fprintf}.
4054 @findex PROFILE_HOOK
4056 A C statement or compound statement to output to @var{file} some assembly
4057 code to call the profiling subroutine @code{mcount} even the target does
4058 not support profiling.
4060 @findex NO_PROFILE_COUNTERS
4061 @item NO_PROFILE_COUNTERS
4062 Define this macro if the @code{mcount} subroutine on your system does
4063 not need a counter variable allocated for each function. This is true
4064 for almost all modern implementations. If you define this macro, you
4065 must not use the @var{labelno} argument to @code{FUNCTION_PROFILER}.
4067 @findex PROFILE_BEFORE_PROLOGUE
4068 @item PROFILE_BEFORE_PROLOGUE
4069 Define this macro if the code for function profiling should come before
4070 the function prologue. Normally, the profiling code comes after.
4073 @findex TARGET_ALLOWS_PROFILING_WITHOUT_FRAME_POINTER
4074 @item TARGET_ALLOWS_PROFILING_WITHOUT_FRAME_POINTER
4075 On some targets, it is impossible to use profiling when the frame
4076 pointer has been omitted. For example, on x86 GNU/Linux systems,
4077 the @code{mcount} routine provided by the GNU C Library finds the
4078 address of the routine that called the routine that called @code{mcount}
4079 by looking in the immediate caller's stack frame. If the immediate
4080 caller has no frame pointer, this lookup will fail.
4082 By default, GCC assumes that the target does allow profiling when the
4083 frame pointer is omitted. This macro should be defined to a C
4084 expression that evaluates to @code{false} if the target does not allow
4085 profiling when the frame pointer is omitted.
4090 @subsection Permitting tail calls
4094 @findex FUNCTION_OK_FOR_SIBCALL
4095 @item FUNCTION_OK_FOR_SIBCALL (@var{decl})
4096 A C expression that evaluates to true if it is ok to perform a sibling
4097 call to @var{decl} from the current function.
4099 It is not uncommon for limitations of calling conventions to prevent
4100 tail calls to functions outside the current unit of translation, or
4101 during PIC compilation. Use this macro to enforce these restrictions,
4102 as the @code{sibcall} md pattern can not fail, or fall over to a
4107 @section Implementing the Varargs Macros
4108 @cindex varargs implementation
4110 GCC comes with an implementation of @code{<varargs.h>} and
4111 @code{<stdarg.h>} that work without change on machines that pass arguments
4112 on the stack. Other machines require their own implementations of
4113 varargs, and the two machine independent header files must have
4114 conditionals to include it.
4116 ISO @code{<stdarg.h>} differs from traditional @code{<varargs.h>} mainly in
4117 the calling convention for @code{va_start}. The traditional
4118 implementation takes just one argument, which is the variable in which
4119 to store the argument pointer. The ISO implementation of
4120 @code{va_start} takes an additional second argument. The user is
4121 supposed to write the last named argument of the function here.
4123 However, @code{va_start} should not use this argument. The way to find
4124 the end of the named arguments is with the built-in functions described
4128 @findex __builtin_saveregs
4129 @item __builtin_saveregs ()
4130 Use this built-in function to save the argument registers in memory so
4131 that the varargs mechanism can access them. Both ISO and traditional
4132 versions of @code{va_start} must use @code{__builtin_saveregs}, unless
4133 you use @code{SETUP_INCOMING_VARARGS} (see below) instead.
4135 On some machines, @code{__builtin_saveregs} is open-coded under the
4136 control of the macro @code{EXPAND_BUILTIN_SAVEREGS}. On other machines,
4137 it calls a routine written in assembler language, found in
4140 Code generated for the call to @code{__builtin_saveregs} appears at the
4141 beginning of the function, as opposed to where the call to
4142 @code{__builtin_saveregs} is written, regardless of what the code is.
4143 This is because the registers must be saved before the function starts
4144 to use them for its own purposes.
4145 @c i rewrote the first sentence above to fix an overfull hbox. --mew
4148 @findex __builtin_args_info
4149 @item __builtin_args_info (@var{category})
4150 Use this built-in function to find the first anonymous arguments in
4153 In general, a machine may have several categories of registers used for
4154 arguments, each for a particular category of data types. (For example,
4155 on some machines, floating-point registers are used for floating-point
4156 arguments while other arguments are passed in the general registers.)
4157 To make non-varargs functions use the proper calling convention, you
4158 have defined the @code{CUMULATIVE_ARGS} data type to record how many
4159 registers in each category have been used so far
4161 @code{__builtin_args_info} accesses the same data structure of type
4162 @code{CUMULATIVE_ARGS} after the ordinary argument layout is finished
4163 with it, with @var{category} specifying which word to access. Thus, the
4164 value indicates the first unused register in a given category.
4166 Normally, you would use @code{__builtin_args_info} in the implementation
4167 of @code{va_start}, accessing each category just once and storing the
4168 value in the @code{va_list} object. This is because @code{va_list} will
4169 have to update the values, and there is no way to alter the
4170 values accessed by @code{__builtin_args_info}.
4172 @findex __builtin_next_arg
4173 @item __builtin_next_arg (@var{lastarg})
4174 This is the equivalent of @code{__builtin_args_info}, for stack
4175 arguments. It returns the address of the first anonymous stack
4176 argument, as type @code{void *}. If @code{ARGS_GROW_DOWNWARD}, it
4177 returns the address of the location above the first anonymous stack
4178 argument. Use it in @code{va_start} to initialize the pointer for
4179 fetching arguments from the stack. Also use it in @code{va_start} to
4180 verify that the second parameter @var{lastarg} is the last named argument
4181 of the current function.
4183 @findex __builtin_classify_type
4184 @item __builtin_classify_type (@var{object})
4185 Since each machine has its own conventions for which data types are
4186 passed in which kind of register, your implementation of @code{va_arg}
4187 has to embody these conventions. The easiest way to categorize the
4188 specified data type is to use @code{__builtin_classify_type} together
4189 with @code{sizeof} and @code{__alignof__}.
4191 @code{__builtin_classify_type} ignores the value of @var{object},
4192 considering only its data type. It returns an integer describing what
4193 kind of type that is---integer, floating, pointer, structure, and so on.
4195 The file @file{typeclass.h} defines an enumeration that you can use to
4196 interpret the values of @code{__builtin_classify_type}.
4199 These machine description macros help implement varargs:
4202 @findex EXPAND_BUILTIN_SAVEREGS
4203 @item EXPAND_BUILTIN_SAVEREGS ()
4204 If defined, is a C expression that produces the machine-specific code
4205 for a call to @code{__builtin_saveregs}. This code will be moved to the
4206 very beginning of the function, before any parameter access are made.
4207 The return value of this function should be an RTX that contains the
4208 value to use as the return of @code{__builtin_saveregs}.
4210 @findex SETUP_INCOMING_VARARGS
4211 @item SETUP_INCOMING_VARARGS (@var{args_so_far}, @var{mode}, @var{type}, @var{pretend_args_size}, @var{second_time})
4212 This macro offers an alternative to using @code{__builtin_saveregs} and
4213 defining the macro @code{EXPAND_BUILTIN_SAVEREGS}. Use it to store the
4214 anonymous register arguments into the stack so that all the arguments
4215 appear to have been passed consecutively on the stack. Once this is
4216 done, you can use the standard implementation of varargs that works for
4217 machines that pass all their arguments on the stack.
4219 The argument @var{args_so_far} is the @code{CUMULATIVE_ARGS} data
4220 structure, containing the values that are obtained after processing the
4221 named arguments. The arguments @var{mode} and @var{type} describe the
4222 last named argument---its machine mode and its data type as a tree node.
4224 The macro implementation should do two things: first, push onto the
4225 stack all the argument registers @emph{not} used for the named
4226 arguments, and second, store the size of the data thus pushed into the
4227 @code{int}-valued variable whose name is supplied as the argument
4228 @var{pretend_args_size}. The value that you store here will serve as
4229 additional offset for setting up the stack frame.
4231 Because you must generate code to push the anonymous arguments at
4232 compile time without knowing their data types,
4233 @code{SETUP_INCOMING_VARARGS} is only useful on machines that have just
4234 a single category of argument register and use it uniformly for all data
4237 If the argument @var{second_time} is nonzero, it means that the
4238 arguments of the function are being analyzed for the second time. This
4239 happens for an inline function, which is not actually compiled until the
4240 end of the source file. The macro @code{SETUP_INCOMING_VARARGS} should
4241 not generate any instructions in this case.
4243 @findex STRICT_ARGUMENT_NAMING
4244 @item STRICT_ARGUMENT_NAMING
4245 Define this macro to be a nonzero value if the location where a function
4246 argument is passed depends on whether or not it is a named argument.
4248 This macro controls how the @var{named} argument to @code{FUNCTION_ARG}
4249 is set for varargs and stdarg functions. If this macro returns a
4250 nonzero value, the @var{named} argument is always true for named
4251 arguments, and false for unnamed arguments. If it returns a value of
4252 zero, but @code{SETUP_INCOMING_VARARGS} is defined, then all arguments
4253 are treated as named. Otherwise, all named arguments except the last
4254 are treated as named.
4256 You need not define this macro if it always returns zero.
4258 @findex PRETEND_OUTGOING_VARARGS_NAMED
4259 @item PRETEND_OUTGOING_VARARGS_NAMED
4260 If you need to conditionally change ABIs so that one works with
4261 @code{SETUP_INCOMING_VARARGS}, but the other works like neither
4262 @code{SETUP_INCOMING_VARARGS} nor @code{STRICT_ARGUMENT_NAMING} was
4263 defined, then define this macro to return nonzero if
4264 @code{SETUP_INCOMING_VARARGS} is used, zero otherwise.
4265 Otherwise, you should not define this macro.
4269 @section Trampolines for Nested Functions
4270 @cindex trampolines for nested functions
4271 @cindex nested functions, trampolines for
4273 A @dfn{trampoline} is a small piece of code that is created at run time
4274 when the address of a nested function is taken. It normally resides on
4275 the stack, in the stack frame of the containing function. These macros
4276 tell GCC how to generate code to allocate and initialize a
4279 The instructions in the trampoline must do two things: load a constant
4280 address into the static chain register, and jump to the real address of
4281 the nested function. On CISC machines such as the m68k, this requires
4282 two instructions, a move immediate and a jump. Then the two addresses
4283 exist in the trampoline as word-long immediate operands. On RISC
4284 machines, it is often necessary to load each address into a register in
4285 two parts. Then pieces of each address form separate immediate
4288 The code generated to initialize the trampoline must store the variable
4289 parts---the static chain value and the function address---into the
4290 immediate operands of the instructions. On a CISC machine, this is
4291 simply a matter of copying each address to a memory reference at the
4292 proper offset from the start of the trampoline. On a RISC machine, it
4293 may be necessary to take out pieces of the address and store them
4297 @findex TRAMPOLINE_TEMPLATE
4298 @item TRAMPOLINE_TEMPLATE (@var{file})
4299 A C statement to output, on the stream @var{file}, assembler code for a
4300 block of data that contains the constant parts of a trampoline. This
4301 code should not include a label---the label is taken care of
4304 If you do not define this macro, it means no template is needed
4305 for the target. Do not define this macro on systems where the block move
4306 code to copy the trampoline into place would be larger than the code
4307 to generate it on the spot.
4309 @findex TRAMPOLINE_SECTION
4310 @item TRAMPOLINE_SECTION
4311 The name of a subroutine to switch to the section in which the
4312 trampoline template is to be placed (@pxref{Sections}). The default is
4313 a value of @samp{readonly_data_section}, which places the trampoline in
4314 the section containing read-only data.
4316 @findex TRAMPOLINE_SIZE
4317 @item TRAMPOLINE_SIZE
4318 A C expression for the size in bytes of the trampoline, as an integer.
4320 @findex TRAMPOLINE_ALIGNMENT
4321 @item TRAMPOLINE_ALIGNMENT
4322 Alignment required for trampolines, in bits.
4324 If you don't define this macro, the value of @code{BIGGEST_ALIGNMENT}
4325 is used for aligning trampolines.
4327 @findex INITIALIZE_TRAMPOLINE
4328 @item INITIALIZE_TRAMPOLINE (@var{addr}, @var{fnaddr}, @var{static_chain})
4329 A C statement to initialize the variable parts of a trampoline.
4330 @var{addr} is an RTX for the address of the trampoline; @var{fnaddr} is
4331 an RTX for the address of the nested function; @var{static_chain} is an
4332 RTX for the static chain value that should be passed to the function
4335 @findex TRAMPOLINE_ADJUST_ADDRESS
4336 @item TRAMPOLINE_ADJUST_ADDRESS (@var{addr})
4337 A C statement that should perform any machine-specific adjustment in
4338 the address of the trampoline. Its argument contains the address that
4339 was passed to @code{INITIALIZE_TRAMPOLINE}. In case the address to be
4340 used for a function call should be different from the address in which
4341 the template was stored, the different address should be assigned to
4342 @var{addr}. If this macro is not defined, @var{addr} will be used for
4345 @findex ALLOCATE_TRAMPOLINE
4346 @item ALLOCATE_TRAMPOLINE (@var{fp})
4347 A C expression to allocate run-time space for a trampoline. The
4348 expression value should be an RTX representing a memory reference to the
4349 space for the trampoline.
4351 @cindex @code{TARGET_ASM_FUNCTION_EPILOGUE} and trampolines
4352 @cindex @code{TARGET_ASM_FUNCTION_PROLOGUE} and trampolines
4353 If this macro is not defined, by default the trampoline is allocated as
4354 a stack slot. This default is right for most machines. The exceptions
4355 are machines where it is impossible to execute instructions in the stack
4356 area. On such machines, you may have to implement a separate stack,
4357 using this macro in conjunction with @code{TARGET_ASM_FUNCTION_PROLOGUE}
4358 and @code{TARGET_ASM_FUNCTION_EPILOGUE}.
4360 @var{fp} points to a data structure, a @code{struct function}, which
4361 describes the compilation status of the immediate containing function of
4362 the function which the trampoline is for. Normally (when
4363 @code{ALLOCATE_TRAMPOLINE} is not defined), the stack slot for the
4364 trampoline is in the stack frame of this containing function. Other
4365 allocation strategies probably must do something analogous with this
4369 Implementing trampolines is difficult on many machines because they have
4370 separate instruction and data caches. Writing into a stack location
4371 fails to clear the memory in the instruction cache, so when the program
4372 jumps to that location, it executes the old contents.
4374 Here are two possible solutions. One is to clear the relevant parts of
4375 the instruction cache whenever a trampoline is set up. The other is to
4376 make all trampolines identical, by having them jump to a standard
4377 subroutine. The former technique makes trampoline execution faster; the
4378 latter makes initialization faster.
4380 To clear the instruction cache when a trampoline is initialized, define
4381 the following macros which describe the shape of the cache.
4384 @findex INSN_CACHE_SIZE
4385 @item INSN_CACHE_SIZE
4386 The total size in bytes of the cache.
4388 @findex INSN_CACHE_LINE_WIDTH
4389 @item INSN_CACHE_LINE_WIDTH
4390 The length in bytes of each cache line. The cache is divided into cache
4391 lines which are disjoint slots, each holding a contiguous chunk of data
4392 fetched from memory. Each time data is brought into the cache, an
4393 entire line is read at once. The data loaded into a cache line is
4394 always aligned on a boundary equal to the line size.
4396 @findex INSN_CACHE_DEPTH
4397 @item INSN_CACHE_DEPTH
4398 The number of alternative cache lines that can hold any particular memory
4402 Alternatively, if the machine has system calls or instructions to clear
4403 the instruction cache directly, you can define the following macro.
4406 @findex CLEAR_INSN_CACHE
4407 @item CLEAR_INSN_CACHE (@var{beg}, @var{end})
4408 If defined, expands to a C expression clearing the @emph{instruction
4409 cache} in the specified interval. If it is not defined, and the macro
4410 @code{INSN_CACHE_SIZE} is defined, some generic code is generated to clear the
4411 cache. The definition of this macro would typically be a series of
4412 @code{asm} statements. Both @var{beg} and @var{end} are both pointer
4416 To use a standard subroutine, define the following macro. In addition,
4417 you must make sure that the instructions in a trampoline fill an entire
4418 cache line with identical instructions, or else ensure that the
4419 beginning of the trampoline code is always aligned at the same point in
4420 its cache line. Look in @file{m68k.h} as a guide.
4423 @findex TRANSFER_FROM_TRAMPOLINE
4424 @item TRANSFER_FROM_TRAMPOLINE
4425 Define this macro if trampolines need a special subroutine to do their
4426 work. The macro should expand to a series of @code{asm} statements
4427 which will be compiled with GCC@. They go in a library function named
4428 @code{__transfer_from_trampoline}.
4430 If you need to avoid executing the ordinary prologue code of a compiled
4431 C function when you jump to the subroutine, you can do so by placing a
4432 special label of your own in the assembler code. Use one @code{asm}
4433 statement to generate an assembler label, and another to make the label
4434 global. Then trampolines can use that label to jump directly to your
4435 special assembler code.
4439 @section Implicit Calls to Library Routines
4440 @cindex library subroutine names
4441 @cindex @file{libgcc.a}
4443 @c prevent bad page break with this line
4444 Here is an explanation of implicit calls to library routines.
4447 @findex MULSI3_LIBCALL
4448 @item MULSI3_LIBCALL
4449 A C string constant giving the name of the function to call for
4450 multiplication of one signed full-word by another. If you do not
4451 define this macro, the default name is used, which is @code{__mulsi3},
4452 a function defined in @file{libgcc.a}.
4454 @findex DIVSI3_LIBCALL
4455 @item DIVSI3_LIBCALL
4456 A C string constant giving the name of the function to call for
4457 division of one signed full-word by another. If you do not define
4458 this macro, the default name is used, which is @code{__divsi3}, a
4459 function defined in @file{libgcc.a}.
4461 @findex UDIVSI3_LIBCALL
4462 @item UDIVSI3_LIBCALL
4463 A C string constant giving the name of the function to call for
4464 division of one unsigned full-word by another. If you do not define
4465 this macro, the default name is used, which is @code{__udivsi3}, a
4466 function defined in @file{libgcc.a}.
4468 @findex MODSI3_LIBCALL
4469 @item MODSI3_LIBCALL
4470 A C string constant giving the name of the function to call for the
4471 remainder in division of one signed full-word by another. If you do
4472 not define this macro, the default name is used, which is
4473 @code{__modsi3}, a function defined in @file{libgcc.a}.
4475 @findex UMODSI3_LIBCALL
4476 @item UMODSI3_LIBCALL
4477 A C string constant giving the name of the function to call for the
4478 remainder in division of one unsigned full-word by another. If you do
4479 not define this macro, the default name is used, which is
4480 @code{__umodsi3}, a function defined in @file{libgcc.a}.
4482 @findex MULDI3_LIBCALL
4483 @item MULDI3_LIBCALL
4484 A C string constant giving the name of the function to call for
4485 multiplication of one signed double-word by another. If you do not
4486 define this macro, the default name is used, which is @code{__muldi3},
4487 a function defined in @file{libgcc.a}.
4489 @findex DIVDI3_LIBCALL
4490 @item DIVDI3_LIBCALL
4491 A C string constant giving the name of the function to call for
4492 division of one signed double-word by another. If you do not define
4493 this macro, the default name is used, which is @code{__divdi3}, a
4494 function defined in @file{libgcc.a}.
4496 @findex UDIVDI3_LIBCALL
4497 @item UDIVDI3_LIBCALL
4498 A C string constant giving the name of the function to call for
4499 division of one unsigned full-word by another. If you do not define
4500 this macro, the default name is used, which is @code{__udivdi3}, a
4501 function defined in @file{libgcc.a}.
4503 @findex MODDI3_LIBCALL
4504 @item MODDI3_LIBCALL
4505 A C string constant giving the name of the function to call for the
4506 remainder in division of one signed double-word by another. If you do
4507 not define this macro, the default name is used, which is
4508 @code{__moddi3}, a function defined in @file{libgcc.a}.
4510 @findex UMODDI3_LIBCALL
4511 @item UMODDI3_LIBCALL
4512 A C string constant giving the name of the function to call for the
4513 remainder in division of one unsigned full-word by another. If you do
4514 not define this macro, the default name is used, which is
4515 @code{__umoddi3}, a function defined in @file{libgcc.a}.
4517 @findex INIT_TARGET_OPTABS
4518 @item INIT_TARGET_OPTABS
4519 Define this macro as a C statement that declares additional library
4520 routines renames existing ones. @code{init_optabs} calls this macro after
4521 initializing all the normal library routines.
4523 @findex FLOAT_LIB_COMPARE_RETURNS_BOOL (@var{mode}, @var{comparison})
4524 @item FLOAT_LIB_COMPARE_RETURNS_BOOL
4525 Define this macro as a C statement that returns nonzero if a call to
4526 the floating point comparison library function will return a boolean
4527 value that indicates the result of the comparison. It should return
4528 zero if one of gcc's own libgcc functions is called.
4530 Most ports don't need to define this macro.
4533 @cindex @code{EDOM}, implicit usage
4535 The value of @code{EDOM} on the target machine, as a C integer constant
4536 expression. If you don't define this macro, GCC does not attempt to
4537 deposit the value of @code{EDOM} into @code{errno} directly. Look in
4538 @file{/usr/include/errno.h} to find the value of @code{EDOM} on your
4541 If you do not define @code{TARGET_EDOM}, then compiled code reports
4542 domain errors by calling the library function and letting it report the
4543 error. If mathematical functions on your system use @code{matherr} when
4544 there is an error, then you should leave @code{TARGET_EDOM} undefined so
4545 that @code{matherr} is used normally.
4547 @findex GEN_ERRNO_RTX
4548 @cindex @code{errno}, implicit usage
4550 Define this macro as a C expression to create an rtl expression that
4551 refers to the global ``variable'' @code{errno}. (On certain systems,
4552 @code{errno} may not actually be a variable.) If you don't define this
4553 macro, a reasonable default is used.
4555 @findex TARGET_MEM_FUNCTIONS
4556 @cindex @code{bcopy}, implicit usage
4557 @cindex @code{memcpy}, implicit usage
4558 @cindex @code{memmove}, implicit usage
4559 @cindex @code{bzero}, implicit usage
4560 @cindex @code{memset}, implicit usage
4561 @item TARGET_MEM_FUNCTIONS
4562 Define this macro if GCC should generate calls to the ISO C
4563 (and System V) library functions @code{memcpy}, @code{memmove} and
4564 @code{memset} rather than the BSD functions @code{bcopy} and @code{bzero}.
4566 @findex LIBGCC_NEEDS_DOUBLE
4567 @item LIBGCC_NEEDS_DOUBLE
4568 Define this macro if @code{float} arguments cannot be passed to library
4569 routines (so they must be converted to @code{double}). This macro
4570 affects both how library calls are generated and how the library
4571 routines in @file{libgcc.a} accept their arguments. It is useful on
4572 machines where floating and fixed point arguments are passed
4573 differently, such as the i860.
4575 @findex NEXT_OBJC_RUNTIME
4576 @item NEXT_OBJC_RUNTIME
4577 Define this macro to generate code for Objective-C message sending using
4578 the calling convention of the NeXT system. This calling convention
4579 involves passing the object, the selector and the method arguments all
4580 at once to the method-lookup library function.
4582 The default calling convention passes just the object and the selector
4583 to the lookup function, which returns a pointer to the method.
4586 @node Addressing Modes
4587 @section Addressing Modes
4588 @cindex addressing modes
4590 @c prevent bad page break with this line
4591 This is about addressing modes.
4594 @findex HAVE_PRE_INCREMENT
4595 @findex HAVE_PRE_DECREMENT
4596 @findex HAVE_POST_INCREMENT
4597 @findex HAVE_POST_DECREMENT
4598 @item HAVE_PRE_INCREMENT
4599 @itemx HAVE_PRE_DECREMENT
4600 @itemx HAVE_POST_INCREMENT
4601 @itemx HAVE_POST_DECREMENT
4602 A C expression that is nonzero if the machine supports pre-increment,
4603 pre-decrement, post-increment, or post-decrement addressing respectively.
4605 @findex HAVE_POST_MODIFY_DISP
4606 @findex HAVE_PRE_MODIFY_DISP
4607 @item HAVE_PRE_MODIFY_DISP
4608 @itemx HAVE_POST_MODIFY_DISP
4609 A C expression that is nonzero if the machine supports pre- or
4610 post-address side-effect generation involving constants other than
4611 the size of the memory operand.
4613 @findex HAVE_POST_MODIFY_REG
4614 @findex HAVE_PRE_MODIFY_REG
4615 @item HAVE_PRE_MODIFY_REG
4616 @itemx HAVE_POST_MODIFY_REG
4617 A C expression that is nonzero if the machine supports pre- or
4618 post-address side-effect generation involving a register displacement.
4620 @findex CONSTANT_ADDRESS_P
4621 @item CONSTANT_ADDRESS_P (@var{x})
4622 A C expression that is 1 if the RTX @var{x} is a constant which
4623 is a valid address. On most machines, this can be defined as
4624 @code{CONSTANT_P (@var{x})}, but a few machines are more restrictive
4625 in which constant addresses are supported.
4628 @code{CONSTANT_P} accepts integer-values expressions whose values are
4629 not explicitly known, such as @code{symbol_ref}, @code{label_ref}, and
4630 @code{high} expressions and @code{const} arithmetic expressions, in
4631 addition to @code{const_int} and @code{const_double} expressions.
4633 @findex MAX_REGS_PER_ADDRESS
4634 @item MAX_REGS_PER_ADDRESS
4635 A number, the maximum number of registers that can appear in a valid
4636 memory address. Note that it is up to you to specify a value equal to
4637 the maximum number that @code{GO_IF_LEGITIMATE_ADDRESS} would ever
4640 @findex GO_IF_LEGITIMATE_ADDRESS
4641 @item GO_IF_LEGITIMATE_ADDRESS (@var{mode}, @var{x}, @var{label})
4642 A C compound statement with a conditional @code{goto @var{label};}
4643 executed if @var{x} (an RTX) is a legitimate memory address on the
4644 target machine for a memory operand of mode @var{mode}.
4646 It usually pays to define several simpler macros to serve as
4647 subroutines for this one. Otherwise it may be too complicated to
4650 This macro must exist in two variants: a strict variant and a
4651 non-strict one. The strict variant is used in the reload pass. It
4652 must be defined so that any pseudo-register that has not been
4653 allocated a hard register is considered a memory reference. In
4654 contexts where some kind of register is required, a pseudo-register
4655 with no hard register must be rejected.
4657 The non-strict variant is used in other passes. It must be defined to
4658 accept all pseudo-registers in every context where some kind of
4659 register is required.
4661 @findex REG_OK_STRICT
4662 Compiler source files that want to use the strict variant of this
4663 macro define the macro @code{REG_OK_STRICT}. You should use an
4664 @code{#ifdef REG_OK_STRICT} conditional to define the strict variant
4665 in that case and the non-strict variant otherwise.
4667 Subroutines to check for acceptable registers for various purposes (one
4668 for base registers, one for index registers, and so on) are typically
4669 among the subroutines used to define @code{GO_IF_LEGITIMATE_ADDRESS}.
4670 Then only these subroutine macros need have two variants; the higher
4671 levels of macros may be the same whether strict or not.
4673 Normally, constant addresses which are the sum of a @code{symbol_ref}
4674 and an integer are stored inside a @code{const} RTX to mark them as
4675 constant. Therefore, there is no need to recognize such sums
4676 specifically as legitimate addresses. Normally you would simply
4677 recognize any @code{const} as legitimate.
4679 Usually @code{PRINT_OPERAND_ADDRESS} is not prepared to handle constant
4680 sums that are not marked with @code{const}. It assumes that a naked
4681 @code{plus} indicates indexing. If so, then you @emph{must} reject such
4682 naked constant sums as illegitimate addresses, so that none of them will
4683 be given to @code{PRINT_OPERAND_ADDRESS}.
4685 @cindex @code{ENCODE_SECTION_INFO} and address validation
4686 On some machines, whether a symbolic address is legitimate depends on
4687 the section that the address refers to. On these machines, define the
4688 macro @code{ENCODE_SECTION_INFO} to store the information into the
4689 @code{symbol_ref}, and then check for it here. When you see a
4690 @code{const}, you will have to look inside it to find the
4691 @code{symbol_ref} in order to determine the section. @xref{Assembler
4694 @findex saveable_obstack
4695 The best way to modify the name string is by adding text to the
4696 beginning, with suitable punctuation to prevent any ambiguity. Allocate
4697 the new name in @code{saveable_obstack}. You will have to modify
4698 @code{ASM_OUTPUT_LABELREF} to remove and decode the added text and
4699 output the name accordingly, and define @code{STRIP_NAME_ENCODING} to
4700 access the original name string.
4702 You can check the information stored here into the @code{symbol_ref} in
4703 the definitions of the macros @code{GO_IF_LEGITIMATE_ADDRESS} and
4704 @code{PRINT_OPERAND_ADDRESS}.
4706 @findex REG_OK_FOR_BASE_P
4707 @item REG_OK_FOR_BASE_P (@var{x})
4708 A C expression that is nonzero if @var{x} (assumed to be a @code{reg}
4709 RTX) is valid for use as a base register. For hard registers, it
4710 should always accept those which the hardware permits and reject the
4711 others. Whether the macro accepts or rejects pseudo registers must be
4712 controlled by @code{REG_OK_STRICT} as described above. This usually
4713 requires two variant definitions, of which @code{REG_OK_STRICT}
4714 controls the one actually used.
4716 @findex REG_MODE_OK_FOR_BASE_P
4717 @item REG_MODE_OK_FOR_BASE_P (@var{x}, @var{mode})
4718 A C expression that is just like @code{REG_OK_FOR_BASE_P}, except that
4719 that expression may examine the mode of the memory reference in
4720 @var{mode}. You should define this macro if the mode of the memory
4721 reference affects whether a register may be used as a base register. If
4722 you define this macro, the compiler will use it instead of
4723 @code{REG_OK_FOR_BASE_P}.
4725 @findex REG_OK_FOR_INDEX_P
4726 @item REG_OK_FOR_INDEX_P (@var{x})
4727 A C expression that is nonzero if @var{x} (assumed to be a @code{reg}
4728 RTX) is valid for use as an index register.
4730 The difference between an index register and a base register is that
4731 the index register may be scaled. If an address involves the sum of
4732 two registers, neither one of them scaled, then either one may be
4733 labeled the ``base'' and the other the ``index''; but whichever
4734 labeling is used must fit the machine's constraints of which registers
4735 may serve in each capacity. The compiler will try both labelings,
4736 looking for one that is valid, and will reload one or both registers
4737 only if neither labeling works.
4739 @findex FIND_BASE_TERM
4740 @item FIND_BASE_TERM (@var{x})
4741 A C expression to determine the base term of address @var{x}.
4742 This macro is used in only one place: `find_base_term' in alias.c.
4744 It is always safe for this macro to not be defined. It exists so
4745 that alias analysis can understand machine-dependent addresses.
4747 The typical use of this macro is to handle addresses containing
4748 a label_ref or symbol_ref within an UNSPEC@.
4750 @findex LEGITIMIZE_ADDRESS
4751 @item LEGITIMIZE_ADDRESS (@var{x}, @var{oldx}, @var{mode}, @var{win})
4752 A C compound statement that attempts to replace @var{x} with a valid
4753 memory address for an operand of mode @var{mode}. @var{win} will be a
4754 C statement label elsewhere in the code; the macro definition may use
4757 GO_IF_LEGITIMATE_ADDRESS (@var{mode}, @var{x}, @var{win});
4761 to avoid further processing if the address has become legitimate.
4763 @findex break_out_memory_refs
4764 @var{x} will always be the result of a call to @code{break_out_memory_refs},
4765 and @var{oldx} will be the operand that was given to that function to produce
4768 The code generated by this macro should not alter the substructure of
4769 @var{x}. If it transforms @var{x} into a more legitimate form, it
4770 should assign @var{x} (which will always be a C variable) a new value.
4772 It is not necessary for this macro to come up with a legitimate
4773 address. The compiler has standard ways of doing so in all cases. In
4774 fact, it is safe for this macro to do nothing. But often a
4775 machine-dependent strategy can generate better code.
4777 @findex LEGITIMIZE_RELOAD_ADDRESS
4778 @item LEGITIMIZE_RELOAD_ADDRESS (@var{x}, @var{mode}, @var{opnum}, @var{type}, @var{ind_levels}, @var{win})
4779 A C compound statement that attempts to replace @var{x}, which is an address
4780 that needs reloading, with a valid memory address for an operand of mode
4781 @var{mode}. @var{win} will be a C statement label elsewhere in the code.
4782 It is not necessary to define this macro, but it might be useful for
4783 performance reasons.
4785 For example, on the i386, it is sometimes possible to use a single
4786 reload register instead of two by reloading a sum of two pseudo
4787 registers into a register. On the other hand, for number of RISC
4788 processors offsets are limited so that often an intermediate address
4789 needs to be generated in order to address a stack slot. By defining
4790 @code{LEGITIMIZE_RELOAD_ADDRESS} appropriately, the intermediate addresses
4791 generated for adjacent some stack slots can be made identical, and thus
4794 @emph{Note}: This macro should be used with caution. It is necessary
4795 to know something of how reload works in order to effectively use this,
4796 and it is quite easy to produce macros that build in too much knowledge
4797 of reload internals.
4799 @emph{Note}: This macro must be able to reload an address created by a
4800 previous invocation of this macro. If it fails to handle such addresses
4801 then the compiler may generate incorrect code or abort.
4804 The macro definition should use @code{push_reload} to indicate parts that
4805 need reloading; @var{opnum}, @var{type} and @var{ind_levels} are usually
4806 suitable to be passed unaltered to @code{push_reload}.
4808 The code generated by this macro must not alter the substructure of
4809 @var{x}. If it transforms @var{x} into a more legitimate form, it
4810 should assign @var{x} (which will always be a C variable) a new value.
4811 This also applies to parts that you change indirectly by calling
4814 @findex strict_memory_address_p
4815 The macro definition may use @code{strict_memory_address_p} to test if
4816 the address has become legitimate.
4819 If you want to change only a part of @var{x}, one standard way of doing
4820 this is to use @code{copy_rtx}. Note, however, that is unshares only a
4821 single level of rtl. Thus, if the part to be changed is not at the
4822 top level, you'll need to replace first the top level.
4823 It is not necessary for this macro to come up with a legitimate
4824 address; but often a machine-dependent strategy can generate better code.
4826 @findex GO_IF_MODE_DEPENDENT_ADDRESS
4827 @item GO_IF_MODE_DEPENDENT_ADDRESS (@var{addr}, @var{label})
4828 A C statement or compound statement with a conditional @code{goto
4829 @var{label};} executed if memory address @var{x} (an RTX) can have
4830 different meanings depending on the machine mode of the memory
4831 reference it is used for or if the address is valid for some modes
4834 Autoincrement and autodecrement addresses typically have mode-dependent
4835 effects because the amount of the increment or decrement is the size
4836 of the operand being addressed. Some machines have other mode-dependent
4837 addresses. Many RISC machines have no mode-dependent addresses.
4839 You may assume that @var{addr} is a valid address for the machine.
4841 @findex LEGITIMATE_CONSTANT_P
4842 @item LEGITIMATE_CONSTANT_P (@var{x})
4843 A C expression that is nonzero if @var{x} is a legitimate constant for
4844 an immediate operand on the target machine. You can assume that
4845 @var{x} satisfies @code{CONSTANT_P}, so you need not check this. In fact,
4846 @samp{1} is a suitable definition for this macro on machines where
4847 anything @code{CONSTANT_P} is valid.
4850 @node Condition Code
4851 @section Condition Code Status
4852 @cindex condition code status
4854 @c prevent bad page break with this line
4855 This describes the condition code status.
4858 The file @file{conditions.h} defines a variable @code{cc_status} to
4859 describe how the condition code was computed (in case the interpretation of
4860 the condition code depends on the instruction that it was set by). This
4861 variable contains the RTL expressions on which the condition code is
4862 currently based, and several standard flags.
4864 Sometimes additional machine-specific flags must be defined in the machine
4865 description header file. It can also add additional machine-specific
4866 information by defining @code{CC_STATUS_MDEP}.
4869 @findex CC_STATUS_MDEP
4870 @item CC_STATUS_MDEP
4871 C code for a data type which is used for declaring the @code{mdep}
4872 component of @code{cc_status}. It defaults to @code{int}.
4874 This macro is not used on machines that do not use @code{cc0}.
4876 @findex CC_STATUS_MDEP_INIT
4877 @item CC_STATUS_MDEP_INIT
4878 A C expression to initialize the @code{mdep} field to ``empty''.
4879 The default definition does nothing, since most machines don't use
4880 the field anyway. If you want to use the field, you should probably
4881 define this macro to initialize it.
4883 This macro is not used on machines that do not use @code{cc0}.
4885 @findex NOTICE_UPDATE_CC
4886 @item NOTICE_UPDATE_CC (@var{exp}, @var{insn})
4887 A C compound statement to set the components of @code{cc_status}
4888 appropriately for an insn @var{insn} whose body is @var{exp}. It is
4889 this macro's responsibility to recognize insns that set the condition
4890 code as a byproduct of other activity as well as those that explicitly
4893 This macro is not used on machines that do not use @code{cc0}.
4895 If there are insns that do not set the condition code but do alter
4896 other machine registers, this macro must check to see whether they
4897 invalidate the expressions that the condition code is recorded as
4898 reflecting. For example, on the 68000, insns that store in address
4899 registers do not set the condition code, which means that usually
4900 @code{NOTICE_UPDATE_CC} can leave @code{cc_status} unaltered for such
4901 insns. But suppose that the previous insn set the condition code
4902 based on location @samp{a4@@(102)} and the current insn stores a new
4903 value in @samp{a4}. Although the condition code is not changed by
4904 this, it will no longer be true that it reflects the contents of
4905 @samp{a4@@(102)}. Therefore, @code{NOTICE_UPDATE_CC} must alter
4906 @code{cc_status} in this case to say that nothing is known about the
4907 condition code value.
4909 The definition of @code{NOTICE_UPDATE_CC} must be prepared to deal
4910 with the results of peephole optimization: insns whose patterns are
4911 @code{parallel} RTXs containing various @code{reg}, @code{mem} or
4912 constants which are just the operands. The RTL structure of these
4913 insns is not sufficient to indicate what the insns actually do. What
4914 @code{NOTICE_UPDATE_CC} should do when it sees one is just to run
4915 @code{CC_STATUS_INIT}.
4917 A possible definition of @code{NOTICE_UPDATE_CC} is to call a function
4918 that looks at an attribute (@pxref{Insn Attributes}) named, for example,
4919 @samp{cc}. This avoids having detailed information about patterns in
4920 two places, the @file{md} file and in @code{NOTICE_UPDATE_CC}.
4922 @findex EXTRA_CC_MODES
4923 @item EXTRA_CC_MODES
4924 A list of additional modes for condition code values in registers
4925 (@pxref{Jump Patterns}). This macro should expand to a sequence of
4926 calls of the macro @code{CC} separated by white space. @code{CC} takes
4927 two arguments. The first is the enumeration name of the mode, which
4928 should begin with @samp{CC} and end with @samp{mode}. The second is a C
4929 string giving the printable name of the mode; it should be the same as
4930 the first argument, but with the trailing @samp{mode} removed.
4932 You should only define this macro if additional modes are required.
4934 A sample definition of @code{EXTRA_CC_MODES} is:
4936 #define EXTRA_CC_MODES \
4937 CC(CC_NOOVmode, "CC_NOOV") \
4938 CC(CCFPmode, "CCFP") \
4939 CC(CCFPEmode, "CCFPE")
4942 @findex SELECT_CC_MODE
4943 @item SELECT_CC_MODE (@var{op}, @var{x}, @var{y})
4944 Returns a mode from class @code{MODE_CC} to be used when comparison
4945 operation code @var{op} is applied to rtx @var{x} and @var{y}. For
4946 example, on the Sparc, @code{SELECT_CC_MODE} is defined as (see
4947 @pxref{Jump Patterns} for a description of the reason for this
4951 #define SELECT_CC_MODE(OP,X,Y) \
4952 (GET_MODE_CLASS (GET_MODE (X)) == MODE_FLOAT \
4953 ? ((OP == EQ || OP == NE) ? CCFPmode : CCFPEmode) \
4954 : ((GET_CODE (X) == PLUS || GET_CODE (X) == MINUS \
4955 || GET_CODE (X) == NEG) \
4956 ? CC_NOOVmode : CCmode))
4959 You need not define this macro if @code{EXTRA_CC_MODES} is not defined.
4961 @findex CANONICALIZE_COMPARISON
4962 @item CANONICALIZE_COMPARISON (@var{code}, @var{op0}, @var{op1})
4963 On some machines not all possible comparisons are defined, but you can
4964 convert an invalid comparison into a valid one. For example, the Alpha
4965 does not have a @code{GT} comparison, but you can use an @code{LT}
4966 comparison instead and swap the order of the operands.
4968 On such machines, define this macro to be a C statement to do any
4969 required conversions. @var{code} is the initial comparison code
4970 and @var{op0} and @var{op1} are the left and right operands of the
4971 comparison, respectively. You should modify @var{code}, @var{op0}, and
4972 @var{op1} as required.
4974 GCC will not assume that the comparison resulting from this macro is
4975 valid but will see if the resulting insn matches a pattern in the
4978 You need not define this macro if it would never change the comparison
4981 @findex REVERSIBLE_CC_MODE
4982 @item REVERSIBLE_CC_MODE (@var{mode})
4983 A C expression whose value is one if it is always safe to reverse a
4984 comparison whose mode is @var{mode}. If @code{SELECT_CC_MODE}
4985 can ever return @var{mode} for a floating-point inequality comparison,
4986 then @code{REVERSIBLE_CC_MODE (@var{mode})} must be zero.
4988 You need not define this macro if it would always returns zero or if the
4989 floating-point format is anything other than @code{IEEE_FLOAT_FORMAT}.
4990 For example, here is the definition used on the Sparc, where floating-point
4991 inequality comparisons are always given @code{CCFPEmode}:
4994 #define REVERSIBLE_CC_MODE(MODE) ((MODE) != CCFPEmode)
4997 @findex REVERSE_CONDITION (@var{code}, @var{mode})
4998 A C expression whose value is reversed condition code of the @var{code} for
4999 comparison done in CC_MODE @var{mode}. The macro is used only in case
5000 @code{REVERSIBLE_CC_MODE (@var{mode})} is nonzero. Define this macro in case
5001 machine has some non-standard way how to reverse certain conditionals. For
5002 instance in case all floating point conditions are non-trapping, compiler may
5003 freely convert unordered compares to ordered one. Then definition may look
5007 #define REVERSE_CONDITION(CODE, MODE) \
5008 ((MODE) != CCFPmode ? reverse_condition (CODE) \
5009 : reverse_condition_maybe_unordered (CODE))
5012 @findex REVERSE_CONDEXEC_PREDICATES_P
5013 @item REVERSE_CONDEXEC_PREDICATES_P (@var{code1}, @var{code2})
5014 A C expression that returns true if the conditional execution predicate
5015 @var{code1} is the inverse of @var{code2} and vice versa. Define this to
5016 return 0 if the target has conditional execution predicates that cannot be
5017 reversed safely. If no expansion is specified, this macro is defined as
5021 #define REVERSE_CONDEXEC_PREDICATES_P (x, y) \
5022 ((x) == reverse_condition (y))
5028 @section Describing Relative Costs of Operations
5029 @cindex costs of instructions
5030 @cindex relative costs
5031 @cindex speed of instructions
5033 These macros let you describe the relative speed of various operations
5034 on the target machine.
5038 @item CONST_COSTS (@var{x}, @var{code}, @var{outer_code})
5039 A part of a C @code{switch} statement that describes the relative costs
5040 of constant RTL expressions. It must contain @code{case} labels for
5041 expression codes @code{const_int}, @code{const}, @code{symbol_ref},
5042 @code{label_ref} and @code{const_double}. Each case must ultimately
5043 reach a @code{return} statement to return the relative cost of the use
5044 of that kind of constant value in an expression. The cost may depend on
5045 the precise value of the constant, which is available for examination in
5046 @var{x}, and the rtx code of the expression in which it is contained,
5047 found in @var{outer_code}.
5049 @var{code} is the expression code---redundant, since it can be
5050 obtained with @code{GET_CODE (@var{x})}.
5053 @findex COSTS_N_INSNS
5054 @item RTX_COSTS (@var{x}, @var{code}, @var{outer_code})
5055 Like @code{CONST_COSTS} but applies to nonconstant RTL expressions.
5056 This can be used, for example, to indicate how costly a multiply
5057 instruction is. In writing this macro, you can use the construct
5058 @code{COSTS_N_INSNS (@var{n})} to specify a cost equal to @var{n} fast
5059 instructions. @var{outer_code} is the code of the expression in which
5060 @var{x} is contained.
5062 This macro is optional; do not define it if the default cost assumptions
5063 are adequate for the target machine.
5065 @findex DEFAULT_RTX_COSTS
5066 @item DEFAULT_RTX_COSTS (@var{x}, @var{code}, @var{outer_code})
5067 This macro, if defined, is called for any case not handled by the
5068 @code{RTX_COSTS} or @code{CONST_COSTS} macros. This eliminates the need
5069 to put case labels into the macro, but the code, or any functions it
5070 calls, must assume that the RTL in @var{x} could be of any type that has
5071 not already been handled. The arguments are the same as for
5072 @code{RTX_COSTS}, and the macro should execute a return statement giving
5073 the cost of any RTL expressions that it can handle. The default cost
5074 calculation is used for any RTL for which this macro does not return a
5077 This macro is optional; do not define it if the default cost assumptions
5078 are adequate for the target machine.
5080 @findex ADDRESS_COST
5081 @item ADDRESS_COST (@var{address})
5082 An expression giving the cost of an addressing mode that contains
5083 @var{address}. If not defined, the cost is computed from
5084 the @var{address} expression and the @code{CONST_COSTS} values.
5086 For most CISC machines, the default cost is a good approximation of the
5087 true cost of the addressing mode. However, on RISC machines, all
5088 instructions normally have the same length and execution time. Hence
5089 all addresses will have equal costs.
5091 In cases where more than one form of an address is known, the form with
5092 the lowest cost will be used. If multiple forms have the same, lowest,
5093 cost, the one that is the most complex will be used.
5095 For example, suppose an address that is equal to the sum of a register
5096 and a constant is used twice in the same basic block. When this macro
5097 is not defined, the address will be computed in a register and memory
5098 references will be indirect through that register. On machines where
5099 the cost of the addressing mode containing the sum is no higher than
5100 that of a simple indirect reference, this will produce an additional
5101 instruction and possibly require an additional register. Proper
5102 specification of this macro eliminates this overhead for such machines.
5104 Similar use of this macro is made in strength reduction of loops.
5106 @var{address} need not be valid as an address. In such a case, the cost
5107 is not relevant and can be any value; invalid addresses need not be
5108 assigned a different cost.
5110 On machines where an address involving more than one register is as
5111 cheap as an address computation involving only one register, defining
5112 @code{ADDRESS_COST} to reflect this can cause two registers to be live
5113 over a region of code where only one would have been if
5114 @code{ADDRESS_COST} were not defined in that manner. This effect should
5115 be considered in the definition of this macro. Equivalent costs should
5116 probably only be given to addresses with different numbers of registers
5117 on machines with lots of registers.
5119 This macro will normally either not be defined or be defined as a
5122 @findex REGISTER_MOVE_COST
5123 @item REGISTER_MOVE_COST (@var{mode}, @var{from}, @var{to})
5124 A C expression for the cost of moving data of mode @var{mode} from a
5125 register in class @var{from} to one in class @var{to}. The classes are
5126 expressed using the enumeration values such as @code{GENERAL_REGS}. A
5127 value of 2 is the default; other values are interpreted relative to
5130 It is not required that the cost always equal 2 when @var{from} is the
5131 same as @var{to}; on some machines it is expensive to move between
5132 registers if they are not general registers.
5134 If reload sees an insn consisting of a single @code{set} between two
5135 hard registers, and if @code{REGISTER_MOVE_COST} applied to their
5136 classes returns a value of 2, reload does not check to ensure that the
5137 constraints of the insn are met. Setting a cost of other than 2 will
5138 allow reload to verify that the constraints are met. You should do this
5139 if the @samp{mov@var{m}} pattern's constraints do not allow such copying.
5141 @findex MEMORY_MOVE_COST
5142 @item MEMORY_MOVE_COST (@var{mode}, @var{class}, @var{in})
5143 A C expression for the cost of moving data of mode @var{mode} between a
5144 register of class @var{class} and memory; @var{in} is zero if the value
5145 is to be written to memory, nonzero if it is to be read in. This cost
5146 is relative to those in @code{REGISTER_MOVE_COST}. If moving between
5147 registers and memory is more expensive than between two registers, you
5148 should define this macro to express the relative cost.
5150 If you do not define this macro, GCC uses a default cost of 4 plus
5151 the cost of copying via a secondary reload register, if one is
5152 needed. If your machine requires a secondary reload register to copy
5153 between memory and a register of @var{class} but the reload mechanism is
5154 more complex than copying via an intermediate, define this macro to
5155 reflect the actual cost of the move.
5157 GCC defines the function @code{memory_move_secondary_cost} if
5158 secondary reloads are needed. It computes the costs due to copying via
5159 a secondary register. If your machine copies from memory using a
5160 secondary register in the conventional way but the default base value of
5161 4 is not correct for your machine, define this macro to add some other
5162 value to the result of that function. The arguments to that function
5163 are the same as to this macro.
5167 A C expression for the cost of a branch instruction. A value of 1 is
5168 the default; other values are interpreted relative to that.
5171 Here are additional macros which do not specify precise relative costs,
5172 but only that certain actions are more expensive than GCC would
5176 @findex SLOW_BYTE_ACCESS
5177 @item SLOW_BYTE_ACCESS
5178 Define this macro as a C expression which is nonzero if accessing less
5179 than a word of memory (i.e.@: a @code{char} or a @code{short}) is no
5180 faster than accessing a word of memory, i.e., if such access
5181 require more than one instruction or if there is no difference in cost
5182 between byte and (aligned) word loads.
5184 When this macro is not defined, the compiler will access a field by
5185 finding the smallest containing object; when it is defined, a fullword
5186 load will be used if alignment permits. Unless bytes accesses are
5187 faster than word accesses, using word accesses is preferable since it
5188 may eliminate subsequent memory access if subsequent accesses occur to
5189 other fields in the same word of the structure, but to different bytes.
5191 @findex SLOW_UNALIGNED_ACCESS
5192 @item SLOW_UNALIGNED_ACCESS (@var{mode}, @var{alignment})
5193 Define this macro to be the value 1 if memory accesses described by the
5194 @var{mode} and @var{alignment} parameters have a cost many times greater
5195 than aligned accesses, for example if they are emulated in a trap
5198 When this macro is nonzero, the compiler will act as if
5199 @code{STRICT_ALIGNMENT} were nonzero when generating code for block
5200 moves. This can cause significantly more instructions to be produced.
5201 Therefore, do not set this macro nonzero if unaligned accesses only add a
5202 cycle or two to the time for a memory access.
5204 If the value of this macro is always zero, it need not be defined. If
5205 this macro is defined, it should produce a nonzero value when
5206 @code{STRICT_ALIGNMENT} is nonzero.
5208 @findex DONT_REDUCE_ADDR
5209 @item DONT_REDUCE_ADDR
5210 Define this macro to inhibit strength reduction of memory addresses.
5211 (On some machines, such strength reduction seems to do harm rather
5216 The threshold of number of scalar memory-to-memory move insns, @emph{below}
5217 which a sequence of insns should be generated instead of a
5218 string move insn or a library call. Increasing the value will always
5219 make code faster, but eventually incurs high cost in increased code size.
5221 Note that on machines where the corresponding move insn is a
5222 @code{define_expand} that emits a sequence of insns, this macro counts
5223 the number of such sequences.
5225 If you don't define this, a reasonable default is used.
5227 @findex MOVE_BY_PIECES_P
5228 @item MOVE_BY_PIECES_P (@var{size}, @var{alignment})
5229 A C expression used to determine whether @code{move_by_pieces} will be used to
5230 copy a chunk of memory, or whether some other block move mechanism
5231 will be used. Defaults to 1 if @code{move_by_pieces_ninsns} returns less
5232 than @code{MOVE_RATIO}.
5234 @findex MOVE_MAX_PIECES
5235 @item MOVE_MAX_PIECES
5236 A C expression used by @code{move_by_pieces} to determine the largest unit
5237 a load or store used to copy memory is. Defaults to @code{MOVE_MAX}.
5239 @findex USE_LOAD_POST_INCREMENT
5240 @item USE_LOAD_POST_INCREMENT (@var{mode})
5241 A C expression used to determine whether a load postincrement is a good
5242 thing to use for a given mode. Defaults to the value of
5243 @code{HAVE_POST_INCREMENT}.
5245 @findex USE_LOAD_POST_DECREMENT
5246 @item USE_LOAD_POST_DECREMENT (@var{mode})
5247 A C expression used to determine whether a load postdecrement is a good
5248 thing to use for a given mode. Defaults to the value of
5249 @code{HAVE_POST_DECREMENT}.
5251 @findex USE_LOAD_PRE_INCREMENT
5252 @item USE_LOAD_PRE_INCREMENT (@var{mode})
5253 A C expression used to determine whether a load preincrement is a good
5254 thing to use for a given mode. Defaults to the value of
5255 @code{HAVE_PRE_INCREMENT}.
5257 @findex USE_LOAD_PRE_DECREMENT
5258 @item USE_LOAD_PRE_DECREMENT (@var{mode})
5259 A C expression used to determine whether a load predecrement is a good
5260 thing to use for a given mode. Defaults to the value of
5261 @code{HAVE_PRE_DECREMENT}.
5263 @findex USE_STORE_POST_INCREMENT
5264 @item USE_STORE_POST_INCREMENT (@var{mode})
5265 A C expression used to determine whether a store postincrement is a good
5266 thing to use for a given mode. Defaults to the value of
5267 @code{HAVE_POST_INCREMENT}.
5269 @findex USE_STORE_POST_DECREMENT
5270 @item USE_STORE_POST_DECREMENT (@var{mode})
5271 A C expression used to determine whether a store postdecrement is a good
5272 thing to use for a given mode. Defaults to the value of
5273 @code{HAVE_POST_DECREMENT}.
5275 @findex USE_STORE_PRE_INCREMENT
5276 @item USE_STORE_PRE_INCREMENT (@var{mode})
5277 This macro is used to determine whether a store preincrement is a good
5278 thing to use for a given mode. Defaults to the value of
5279 @code{HAVE_PRE_INCREMENT}.
5281 @findex USE_STORE_PRE_DECREMENT
5282 @item USE_STORE_PRE_DECREMENT (@var{mode})
5283 This macro is used to determine whether a store predecrement is a good
5284 thing to use for a given mode. Defaults to the value of
5285 @code{HAVE_PRE_DECREMENT}.
5287 @findex NO_FUNCTION_CSE
5288 @item NO_FUNCTION_CSE
5289 Define this macro if it is as good or better to call a constant
5290 function address than to call an address kept in a register.
5292 @findex NO_RECURSIVE_FUNCTION_CSE
5293 @item NO_RECURSIVE_FUNCTION_CSE
5294 Define this macro if it is as good or better for a function to call
5295 itself with an explicit address than to call an address kept in a
5300 @section Adjusting the Instruction Scheduler
5302 The instruction scheduler may need a fair amount of machine-specific
5303 adjustment in order to produce good code. GCC provides several target
5304 hooks for this purpose. It is usually enough to define just a few of
5305 them: try the first ones in this list first.
5307 @deftypefn {Target Hook} int TARGET_SCHED_ISSUE_RATE (void)
5308 This hook returns the maximum number of instructions that can ever issue
5309 at the same time on the target machine. The default is one. This value
5310 must be constant over the entire compilation. If you need it to vary
5311 depending on what the instructions are, you must use
5312 @samp{TARGET_SCHED_VARIABLE_ISSUE}.
5315 @deftypefn {Target Hook} int TARGET_SCHED_VARIABLE_ISSUE (FILE *@var{file}, int @var{verbose}, rtx @var{insn}, int @var{more})
5316 This hook is executed by the scheduler after it has scheduled an insn
5317 from the ready list. It should return the number of insns which can
5318 still be issued in the current cycle. Normally this is
5319 @samp{@w{@var{more} - 1}}. You should define this hook if some insns
5320 take more machine resources than others, so that fewer insns can follow
5321 them in the same cycle. @var{file} is either a null pointer, or a stdio
5322 stream to write any debug output to. @var{verbose} is the verbose level
5323 provided by @option{-fsched-verbose-@var{n}}. @var{insn} is the
5324 instruction that was scheduled.
5327 @deftypefn {Target Hook} int TARGET_SCHED_ADJUST_COST (rtx @var{insn}, rtx @var{link}, rtx @var{dep_insn}, int @var{cost})
5328 This function corrects the value of @var{cost} based on the relationship
5329 between @var{insn} and @var{dep_insn} through the dependence @var{link}.
5330 It should return the new value. The default is to make no adjustment to
5331 @var{cost}. This can be used for example to specify to the scheduler
5332 that an output- or anti-dependence does not incur the same cost as a
5336 @deftypefn {Target Hook} int TARGET_SCHED_ADJUST_PRIORITY (rtx @var{insn}, int @var{priority})
5337 This hook adjusts the integer scheduling priority @var{priority} of
5338 @var{insn}. It should return the new priority. Reduce the priority to
5339 execute @var{insn} earlier, increase the priority to execute @var{insn}
5340 later. Do not define this hook if you do not need to adjust the
5341 scheduling priorities of insns.
5344 @deftypefn {Target Hook} int TARGET_SCHED_REORDER (FILE *@var{file}, int @var{verbose}, rtx *@var{ready}, int *@var{n_readyp}, int @var{clock})
5345 This hook is executed by the scheduler after it has scheduled the ready
5346 list, to allow the machine description to reorder it (for example to
5347 combine two small instructions together on @samp{VLIW} machines).
5348 @var{file} is either a null pointer, or a stdio stream to write any
5349 debug output to. @var{verbose} is the verbose level provided by
5350 @option{-fsched-verbose-@var{n}}. @var{ready} is a pointer to the ready
5351 list of instructions that are ready to be scheduled. @var{n_readyp} is
5352 a pointer to the number of elements in the ready list. The scheduler
5353 reads the ready list in reverse order, starting with
5354 @var{ready}[@var{*n_readyp}-1] and going to @var{ready}[0]. @var{clock}
5355 is the timer tick of the scheduler. You may modify the ready list and
5356 the number of ready insns. The return value is the number of insns that
5357 can issue this cycle; normally this is just @code{issue_rate}. See also
5358 @samp{TARGET_SCHED_REORDER2}.
5361 @deftypefn {Target Hook} int TARGET_SCHED_REORDER2 (FILE *@var{file}, int @var{verbose}, rtx *@var{ready}, int *@var{n_ready}, @var{clock})
5362 Like @samp{TARGET_SCHED_REORDER}, but called at a different time. That
5363 function is called whenever the scheduler starts a new cycle. This one
5364 is called once per iteration over a cycle, immediately after
5365 @samp{TARGET_SCHED_VARIABLE_ISSUE}; it can reorder the ready list and
5366 return the number of insns to be scheduled in the same cycle. Defining
5367 this hook can be useful if there are frequent situations where
5368 scheduling one insn causes other insns to become ready in the same
5369 cycle. These other insns can then be taken into account properly.
5372 @deftypefn {Target Hook} void TARGET_SCHED_INIT (FILE *@var{file}, int @var{verbose}, int @var{max_ready})
5373 This hook is executed by the scheduler at the beginning of each block of
5374 instructions that are to be scheduled. @var{file} is either a null
5375 pointer, or a stdio stream to write any debug output to. @var{verbose}
5376 is the verbose level provided by @option{-fsched-verbose-@var{n}}.
5377 @var{max_ready} is the maximum number of insns in the current scheduling
5378 region that can be live at the same time. This can be used to allocate
5379 scratch space if it is needed, e.g. by @samp{TARGET_SCHED_REORDER}.
5382 @deftypefn {Target Hook} void TARGET_SCHED_FINISH (FILE *@var{file}, int @var{verbose})
5383 This hook is executed by the scheduler at the end of each block of
5384 instructions that are to be scheduled. It can be used to perform
5385 cleanup of any actions done by the other scheduling hooks. @var{file}
5386 is either a null pointer, or a stdio stream to write any debug output
5387 to. @var{verbose} is the verbose level provided by
5388 @option{-fsched-verbose-@var{n}}.
5391 @deftypefn {Target Hook} rtx TARGET_SCHED_CYCLE_DISPLAY (int @var{clock}, rtx @var{last})
5392 This hook is called in verbose mode only, at the beginning of each pass
5393 over a basic block. It should insert an insn into the chain after
5394 @var{last}, which has no effect, but records the value @var{clock} in
5395 RTL dumps and assembly output. Define this hook only if you need this
5396 level of detail about what the scheduler is doing.
5400 @section Dividing the Output into Sections (Texts, Data, @dots{})
5401 @c the above section title is WAY too long. maybe cut the part between
5402 @c the (...)? --mew 10feb93
5404 An object file is divided into sections containing different types of
5405 data. In the most common case, there are three sections: the @dfn{text
5406 section}, which holds instructions and read-only data; the @dfn{data
5407 section}, which holds initialized writable data; and the @dfn{bss
5408 section}, which holds uninitialized data. Some systems have other kinds
5411 The compiler must tell the assembler when to switch sections. These
5412 macros control what commands to output to tell the assembler this. You
5413 can also define additional sections.
5416 @findex TEXT_SECTION_ASM_OP
5417 @item TEXT_SECTION_ASM_OP
5418 A C expression whose value is a string, including spacing, containing the
5419 assembler operation that should precede instructions and read-only data.
5420 Normally @code{"\t.text"} is right.
5422 @findex TEXT_SECTION
5424 A C statement that switches to the default section containing instructions.
5425 Normally this is not needed, as simply defining @code{TEXT_SECTION_ASM_OP}
5426 is enough. The MIPS port uses this to sort all functions after all data
5429 @findex DATA_SECTION_ASM_OP
5430 @item DATA_SECTION_ASM_OP
5431 A C expression whose value is a string, including spacing, containing the
5432 assembler operation to identify the following data as writable initialized
5433 data. Normally @code{"\t.data"} is right.
5435 @findex SHARED_SECTION_ASM_OP
5436 @item SHARED_SECTION_ASM_OP
5437 If defined, a C expression whose value is a string, including spacing,
5438 containing the assembler operation to identify the following data as
5439 shared data. If not defined, @code{DATA_SECTION_ASM_OP} will be used.
5441 @findex BSS_SECTION_ASM_OP
5442 @item BSS_SECTION_ASM_OP
5443 If defined, a C expression whose value is a string, including spacing,
5444 containing the assembler operation to identify the following data as
5445 uninitialized global data. If not defined, and neither
5446 @code{ASM_OUTPUT_BSS} nor @code{ASM_OUTPUT_ALIGNED_BSS} are defined,
5447 uninitialized global data will be output in the data section if
5448 @option{-fno-common} is passed, otherwise @code{ASM_OUTPUT_COMMON} will be
5451 @findex SHARED_BSS_SECTION_ASM_OP
5452 @item SHARED_BSS_SECTION_ASM_OP
5453 If defined, a C expression whose value is a string, including spacing,
5454 containing the assembler operation to identify the following data as
5455 uninitialized global shared data. If not defined, and
5456 @code{BSS_SECTION_ASM_OP} is, the latter will be used.
5458 @findex INIT_SECTION_ASM_OP
5459 @item INIT_SECTION_ASM_OP
5460 If defined, a C expression whose value is a string, including spacing,
5461 containing the assembler operation to identify the following data as
5462 initialization code. If not defined, GCC will assume such a section does
5465 @findex FINI_SECTION_ASM_OP
5466 @item FINI_SECTION_ASM_OP
5467 If defined, a C expression whose value is a string, including spacing,
5468 containing the assembler operation to identify the following data as
5469 finalization code. If not defined, GCC will assume such a section does
5472 @findex CRT_CALL_STATIC_FUNCTION
5473 @item CRT_CALL_STATIC_FUNCTION
5474 If defined, a C statement that calls the function named as the sole
5475 argument of this macro. This is used in @file{crtstuff.c} if
5476 @code{INIT_SECTION_ASM_OP} or @code{FINI_SECTION_ASM_OP} to calls to
5477 initialization and finalization functions from the init and fini
5478 sections. By default, this macro is a simple function call. Some
5479 ports need hand-crafted assembly code to avoid dependencies on
5480 registers initialized in the function prologue or to ensure that
5481 constant pools don't end up too far way in the text section.
5483 @findex EXTRA_SECTIONS
5486 @item EXTRA_SECTIONS
5487 A list of names for sections other than the standard two, which are
5488 @code{in_text} and @code{in_data}. You need not define this macro
5489 on a system with no other sections (that GCC needs to use).
5491 @findex EXTRA_SECTION_FUNCTIONS
5492 @findex text_section
5493 @findex data_section
5494 @item EXTRA_SECTION_FUNCTIONS
5495 One or more functions to be defined in @file{varasm.c}. These
5496 functions should do jobs analogous to those of @code{text_section} and
5497 @code{data_section}, for your additional sections. Do not define this
5498 macro if you do not define @code{EXTRA_SECTIONS}.
5500 @findex READONLY_DATA_SECTION
5501 @item READONLY_DATA_SECTION
5502 On most machines, read-only variables, constants, and jump tables are
5503 placed in the text section. If this is not the case on your machine,
5504 this macro should be defined to be the name of a function (either
5505 @code{data_section} or a function defined in @code{EXTRA_SECTIONS}) that
5506 switches to the section to be used for read-only items.
5508 If these items should be placed in the text section, this macro should
5511 @findex SELECT_SECTION
5512 @item SELECT_SECTION (@var{exp}, @var{reloc}, @var{align})
5513 A C statement or statements to switch to the appropriate section for
5514 output of @var{exp}. You can assume that @var{exp} is either a
5515 @code{VAR_DECL} node or a constant of some sort. @var{reloc}
5516 indicates whether the initial value of @var{exp} requires link-time
5517 relocations. Bit 1 is set when variable contains local relocations
5518 only, while bit 2 is set for global relocations.
5519 Select the section by calling @code{text_section} or one
5520 of the alternatives for other sections. @var{align} is the constant
5523 Do not define this macro if you put all read-only variables and
5524 constants in the read-only data section (usually the text section).
5526 @findex SELECT_RTX_SECTION
5527 @item SELECT_RTX_SECTION (@var{mode}, @var{rtx}, @var{align})
5528 A C statement or statements to switch to the appropriate section for
5529 output of @var{rtx} in mode @var{mode}. You can assume that @var{rtx}
5530 is some kind of constant in RTL@. The argument @var{mode} is redundant
5531 except in the case of a @code{const_int} rtx. Select the section by
5532 calling @code{text_section} or one of the alternatives for other
5533 sections. @var{align} is the constant alignment in bits.
5535 Do not define this macro if you put all constants in the read-only
5538 @findex JUMP_TABLES_IN_TEXT_SECTION
5539 @item JUMP_TABLES_IN_TEXT_SECTION
5540 Define this macro to be an expression with a nonzero value if jump
5541 tables (for @code{tablejump} insns) should be output in the text
5542 section, along with the assembler instructions. Otherwise, the
5543 readonly data section is used.
5545 This macro is irrelevant if there is no separate readonly data section.
5547 @findex ENCODE_SECTION_INFO
5548 @item ENCODE_SECTION_INFO (@var{decl})
5549 Define this macro if references to a symbol or a constant must be
5550 treated differently depending on something about the variable or
5551 function named by the symbol (such as what section it is in).
5553 The macro definition, if any, is executed under two circumstances. One
5554 is immediately after the rtl for @var{decl} that represents a variable
5555 or a function has been created and stored in @code{DECL_RTL
5556 (@var{decl})}. The value of the rtl will be a @code{mem} whose address
5557 is a @code{symbol_ref}. The other is immediately after the rtl for
5558 @var{decl} that represents a constant has been created and stored in
5559 @code{TREE_CST_RTL (@var{decl})}. The macro is called once for each
5560 distinct constant in a source file.
5562 @cindex @code{SYMBOL_REF_FLAG}, in @code{ENCODE_SECTION_INFO}
5563 The usual thing for this macro to do is to record a flag in the
5564 @code{symbol_ref} (such as @code{SYMBOL_REF_FLAG}) or to store a
5565 modified name string in the @code{symbol_ref} (if one bit is not enough
5568 @findex STRIP_NAME_ENCODING
5569 @item STRIP_NAME_ENCODING (@var{var}, @var{sym_name})
5570 Decode @var{sym_name} and store the real name part in @var{var}, sans
5571 the characters that encode section info. Define this macro if
5572 @code{ENCODE_SECTION_INFO} alters the symbol's name string.
5574 @findex UNIQUE_SECTION
5575 @item UNIQUE_SECTION (@var{decl}, @var{reloc})
5576 A C statement to build up a unique section name, expressed as a
5577 @code{STRING_CST} node, and assign it to @samp{DECL_SECTION_NAME (@var{decl})}.
5578 @var{reloc} indicates whether the initial value of @var{exp} requires
5579 link-time relocations. If you do not define this macro, GCC will use
5580 the symbol name prefixed by @samp{.} as the section name. Note - this
5581 macro can now be called for uninitialized data items as well as
5582 initialized data and functions.
5586 @section Position Independent Code
5587 @cindex position independent code
5590 This section describes macros that help implement generation of position
5591 independent code. Simply defining these macros is not enough to
5592 generate valid PIC; you must also add support to the macros
5593 @code{GO_IF_LEGITIMATE_ADDRESS} and @code{PRINT_OPERAND_ADDRESS}, as
5594 well as @code{LEGITIMIZE_ADDRESS}. You must modify the definition of
5595 @samp{movsi} to do something appropriate when the source operand
5596 contains a symbolic address. You may also need to alter the handling of
5597 switch statements so that they use relative addresses.
5598 @c i rearranged the order of the macros above to try to force one of
5599 @c them to the next line, to eliminate an overfull hbox. --mew 10feb93
5602 @findex PIC_OFFSET_TABLE_REGNUM
5603 @item PIC_OFFSET_TABLE_REGNUM
5604 The register number of the register used to address a table of static
5605 data addresses in memory. In some cases this register is defined by a
5606 processor's ``application binary interface'' (ABI)@. When this macro
5607 is defined, RTL is generated for this register once, as with the stack
5608 pointer and frame pointer registers. If this macro is not defined, it
5609 is up to the machine-dependent files to allocate such a register (if
5612 @findex PIC_OFFSET_TABLE_REG_CALL_CLOBBERED
5613 @item PIC_OFFSET_TABLE_REG_CALL_CLOBBERED
5614 Define this macro if the register defined by
5615 @code{PIC_OFFSET_TABLE_REGNUM} is clobbered by calls. Do not define
5616 this macro if @code{PIC_OFFSET_TABLE_REGNUM} is not defined.
5618 @findex FINALIZE_PIC
5620 By generating position-independent code, when two different programs (A
5621 and B) share a common library (libC.a), the text of the library can be
5622 shared whether or not the library is linked at the same address for both
5623 programs. In some of these environments, position-independent code
5624 requires not only the use of different addressing modes, but also
5625 special code to enable the use of these addressing modes.
5627 The @code{FINALIZE_PIC} macro serves as a hook to emit these special
5628 codes once the function is being compiled into assembly code, but not
5629 before. (It is not done before, because in the case of compiling an
5630 inline function, it would lead to multiple PIC prologues being
5631 included in functions which used inline functions and were compiled to
5634 @findex LEGITIMATE_PIC_OPERAND_P
5635 @item LEGITIMATE_PIC_OPERAND_P (@var{x})
5636 A C expression that is nonzero if @var{x} is a legitimate immediate
5637 operand on the target machine when generating position independent code.
5638 You can assume that @var{x} satisfies @code{CONSTANT_P}, so you need not
5639 check this. You can also assume @var{flag_pic} is true, so you need not
5640 check it either. You need not define this macro if all constants
5641 (including @code{SYMBOL_REF}) can be immediate operands when generating
5642 position independent code.
5645 @node Assembler Format
5646 @section Defining the Output Assembler Language
5648 This section describes macros whose principal purpose is to describe how
5649 to write instructions in assembler language---rather than what the
5653 * File Framework:: Structural information for the assembler file.
5654 * Data Output:: Output of constants (numbers, strings, addresses).
5655 * Uninitialized Data:: Output of uninitialized variables.
5656 * Label Output:: Output and generation of labels.
5657 * Initialization:: General principles of initialization
5658 and termination routines.
5659 * Macros for Initialization::
5660 Specific macros that control the handling of
5661 initialization and termination routines.
5662 * Instruction Output:: Output of actual instructions.
5663 * Dispatch Tables:: Output of jump tables.
5664 * Exception Region Output:: Output of exception region code.
5665 * Alignment Output:: Pseudo ops for alignment and skipping data.
5668 @node File Framework
5669 @subsection The Overall Framework of an Assembler File
5670 @cindex assembler format
5671 @cindex output of assembler code
5673 @c prevent bad page break with this line
5674 This describes the overall framework of an assembler file.
5677 @findex ASM_FILE_START
5678 @item ASM_FILE_START (@var{stream})
5679 A C expression which outputs to the stdio stream @var{stream}
5680 some appropriate text to go at the start of an assembler file.
5682 Normally this macro is defined to output a line containing
5683 @samp{#NO_APP}, which is a comment that has no effect on most
5684 assemblers but tells the GNU assembler that it can save time by not
5685 checking for certain assembler constructs.
5687 On systems that use SDB, it is necessary to output certain commands;
5688 see @file{attasm.h}.
5690 @findex ASM_FILE_END
5691 @item ASM_FILE_END (@var{stream})
5692 A C expression which outputs to the stdio stream @var{stream}
5693 some appropriate text to go at the end of an assembler file.
5695 If this macro is not defined, the default is to output nothing
5696 special at the end of the file. Most systems don't require any
5699 On systems that use SDB, it is necessary to output certain commands;
5700 see @file{attasm.h}.
5702 @findex ASM_COMMENT_START
5703 @item ASM_COMMENT_START
5704 A C string constant describing how to begin a comment in the target
5705 assembler language. The compiler assumes that the comment will end at
5706 the end of the line.
5710 A C string constant for text to be output before each @code{asm}
5711 statement or group of consecutive ones. Normally this is
5712 @code{"#APP"}, which is a comment that has no effect on most
5713 assemblers but tells the GNU assembler that it must check the lines
5714 that follow for all valid assembler constructs.
5718 A C string constant for text to be output after each @code{asm}
5719 statement or group of consecutive ones. Normally this is
5720 @code{"#NO_APP"}, which tells the GNU assembler to resume making the
5721 time-saving assumptions that are valid for ordinary compiler output.
5723 @findex ASM_OUTPUT_SOURCE_FILENAME
5724 @item ASM_OUTPUT_SOURCE_FILENAME (@var{stream}, @var{name})
5725 A C statement to output COFF information or DWARF debugging information
5726 which indicates that filename @var{name} is the current source file to
5727 the stdio stream @var{stream}.
5729 This macro need not be defined if the standard form of output
5730 for the file format in use is appropriate.
5732 @findex OUTPUT_QUOTED_STRING
5733 @item OUTPUT_QUOTED_STRING (@var{stream}, @var{string})
5734 A C statement to output the string @var{string} to the stdio stream
5735 @var{stream}. If you do not call the function @code{output_quoted_string}
5736 in your config files, GCC will only call it to output filenames to
5737 the assembler source. So you can use it to canonicalize the format
5738 of the filename using this macro.
5740 @findex ASM_OUTPUT_SOURCE_LINE
5741 @item ASM_OUTPUT_SOURCE_LINE (@var{stream}, @var{line})
5742 A C statement to output DBX or SDB debugging information before code
5743 for line number @var{line} of the current source file to the
5744 stdio stream @var{stream}.
5746 This macro need not be defined if the standard form of debugging
5747 information for the debugger in use is appropriate.
5749 @findex ASM_OUTPUT_IDENT
5750 @item ASM_OUTPUT_IDENT (@var{stream}, @var{string})
5751 A C statement to output something to the assembler file to handle a
5752 @samp{#ident} directive containing the text @var{string}. If this
5753 macro is not defined, nothing is output for a @samp{#ident} directive.
5755 @findex OBJC_PROLOGUE
5757 A C statement to output any assembler statements which are required to
5758 precede any Objective-C object definitions or message sending. The
5759 statement is executed only when compiling an Objective-C program.
5762 @deftypefn {Target Hook} void TARGET_ASM_NAMED_SECTION (const char *@var{name}, unsigned int @var{flags}, unsigned int @var{align})
5763 Output assembly directives to switch to section @var{name}. The section
5764 should have attributes as specified by @var{flags}, which is a bit mask
5765 of the @code{SECTION_*} flags defined in @file{output.h}. If @var{align}
5766 is nonzero, it contains an alignment in bytes to be used for the section,
5767 otherwise some target default should be used. Only targets that must
5768 specify an alignment within the section directive need pay attention to
5769 @var{align} -- we will still use @code{ASM_OUTPUT_ALIGN}.
5772 @deftypefn {Target Hook} bool TARGET_HAVE_NAMED_SECTIONS
5773 This flag is true if the target supports @code{TARGET_ASM_NAMED_SECTION}.
5776 @deftypefn {Target Hook} {unsigned int} TARGET_SECTION_TYPE_FLAGS (tree @var{decl}, const char *@var{name}, int @var{reloc})
5777 Choose a set of section attributes for use by @code{TARGET_ASM_NAMED_SECTION}
5778 based on a variable or function decl, a section name, and whether or not the
5779 declaration's initializer may contain runtime relocations. @var{decl} may be
5780 null, in which case read-write data should be assumed.
5782 The default version if this function handles choosing code vs data,
5783 read-only vs read-write data, and @code{flag_pic}. You should only
5784 need to override this if your target has special flags that might be
5785 set via @code{__attribute__}.
5790 @subsection Output of Data
5793 @deftypevr {Target Hook} {const char *} TARGET_ASM_BYTE_OP
5794 @deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_HI_OP
5795 @deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_SI_OP
5796 @deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_DI_OP
5797 @deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_TI_OP
5798 @deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_HI_OP
5799 @deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_SI_OP
5800 @deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_DI_OP
5801 @deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_TI_OP
5802 These hooks specify assembly directives for creating certain kinds
5803 of integer object. The @code{TARGET_ASM_BYTE_OP} directive creates a
5804 byte-sized object, the @code{TARGET_ASM_ALIGNED_HI_OP} one creates an
5805 aligned two-byte object, and so on. Any of the hooks may be
5806 @code{NULL}, indicating that no suitable directive is available.
5808 The compiler will print these strings at the start of a new line,
5809 followed immediately by the object's initial value. In most cases,
5810 the string should contain a tab, a pseudo-op, and then another tab.
5813 @deftypefn {Target Hook} bool TARGET_ASM_INTEGER (rtx @var{x}, unsigned int @var{size}, int @var{aligned_p})
5814 The @code{assemble_integer} function uses this hook to output an
5815 integer object. @var{x} is the object's value, @var{size} is its size
5816 in bytes and @var{aligned_p} indicates whether it is aligned. The
5817 function should return @code{true} if it was able to output the
5818 object. If it returns false, @code{assemble_integer} will try to
5819 split the object into smaller parts.
5821 The default implementation of this hook will use the
5822 @code{TARGET_ASM_BYTE_OP} family of strings, returning @code{false}
5823 when the relevant string is @code{NULL}.
5827 @findex OUTPUT_ADDR_CONST_EXTRA
5828 @item OUTPUT_ADDR_CONST_EXTRA (@var{stream}, @var{x}, @var{fail})
5829 A C statement to recognize @var{rtx} patterns that
5830 @code{output_addr_const} can't deal with, and output assembly code to
5831 @var{stream} corresponding to the pattern @var{x}. This may be used to
5832 allow machine-dependent @code{UNSPEC}s to appear within constants.
5834 If @code{OUTPUT_ADDR_CONST_EXTRA} fails to recognize a pattern, it must
5835 @code{goto fail}, so that a standard error message is printed. If it
5836 prints an error message itself, by calling, for example,
5837 @code{output_operand_lossage}, it may just complete normally.
5839 @findex ASM_OUTPUT_ASCII
5840 @item ASM_OUTPUT_ASCII (@var{stream}, @var{ptr}, @var{len})
5841 A C statement to output to the stdio stream @var{stream} an assembler
5842 instruction to assemble a string constant containing the @var{len}
5843 bytes at @var{ptr}. @var{ptr} will be a C expression of type
5844 @code{char *} and @var{len} a C expression of type @code{int}.
5846 If the assembler has a @code{.ascii} pseudo-op as found in the
5847 Berkeley Unix assembler, do not define the macro
5848 @code{ASM_OUTPUT_ASCII}.
5850 @findex ASM_OUTPUT_FDESC
5851 @item ASM_OUTPUT_FDESC (@var{stream}, @var{decl}, @var{n})
5852 A C statement to output word @var{n} of a function descriptor for
5853 @var{decl}. This must be defined if @code{TARGET_VTABLE_USES_DESCRIPTORS}
5854 is defined, and is otherwise unused.
5856 @findex CONSTANT_POOL_BEFORE_FUNCTION
5857 @item CONSTANT_POOL_BEFORE_FUNCTION
5858 You may define this macro as a C expression. You should define the
5859 expression to have a nonzero value if GCC should output the constant
5860 pool for a function before the code for the function, or a zero value if
5861 GCC should output the constant pool after the function. If you do
5862 not define this macro, the usual case, GCC will output the constant
5863 pool before the function.
5865 @findex ASM_OUTPUT_POOL_PROLOGUE
5866 @item ASM_OUTPUT_POOL_PROLOGUE (@var{file}, @var{funname}, @var{fundecl}, @var{size})
5867 A C statement to output assembler commands to define the start of the
5868 constant pool for a function. @var{funname} is a string giving
5869 the name of the function. Should the return type of the function
5870 be required, it can be obtained via @var{fundecl}. @var{size}
5871 is the size, in bytes, of the constant pool that will be written
5872 immediately after this call.
5874 If no constant-pool prefix is required, the usual case, this macro need
5877 @findex ASM_OUTPUT_SPECIAL_POOL_ENTRY
5878 @item ASM_OUTPUT_SPECIAL_POOL_ENTRY (@var{file}, @var{x}, @var{mode}, @var{align}, @var{labelno}, @var{jumpto})
5879 A C statement (with or without semicolon) to output a constant in the
5880 constant pool, if it needs special treatment. (This macro need not do
5881 anything for RTL expressions that can be output normally.)
5883 The argument @var{file} is the standard I/O stream to output the
5884 assembler code on. @var{x} is the RTL expression for the constant to
5885 output, and @var{mode} is the machine mode (in case @var{x} is a
5886 @samp{const_int}). @var{align} is the required alignment for the value
5887 @var{x}; you should output an assembler directive to force this much
5890 The argument @var{labelno} is a number to use in an internal label for
5891 the address of this pool entry. The definition of this macro is
5892 responsible for outputting the label definition at the proper place.
5893 Here is how to do this:
5896 ASM_OUTPUT_INTERNAL_LABEL (@var{file}, "LC", @var{labelno});
5899 When you output a pool entry specially, you should end with a
5900 @code{goto} to the label @var{jumpto}. This will prevent the same pool
5901 entry from being output a second time in the usual manner.
5903 You need not define this macro if it would do nothing.
5905 @findex CONSTANT_AFTER_FUNCTION_P
5906 @item CONSTANT_AFTER_FUNCTION_P (@var{exp})
5907 Define this macro as a C expression which is nonzero if the constant
5908 @var{exp}, of type @code{tree}, should be output after the code for a
5909 function. The compiler will normally output all constants before the
5910 function; you need not define this macro if this is OK@.
5912 @findex ASM_OUTPUT_POOL_EPILOGUE
5913 @item ASM_OUTPUT_POOL_EPILOGUE (@var{file} @var{funname} @var{fundecl} @var{size})
5914 A C statement to output assembler commands to at the end of the constant
5915 pool for a function. @var{funname} is a string giving the name of the
5916 function. Should the return type of the function be required, you can
5917 obtain it via @var{fundecl}. @var{size} is the size, in bytes, of the
5918 constant pool that GCC wrote immediately before this call.
5920 If no constant-pool epilogue is required, the usual case, you need not
5923 @findex IS_ASM_LOGICAL_LINE_SEPARATOR
5924 @item IS_ASM_LOGICAL_LINE_SEPARATOR (@var{C})
5925 Define this macro as a C expression which is nonzero if @var{C} is
5926 used as a logical line separator by the assembler.
5928 If you do not define this macro, the default is that only
5929 the character @samp{;} is treated as a logical line separator.
5932 @deftypevr {Target Hook} {const char *} TARGET_ASM_OPEN_PAREN
5933 @deftypevrx {Target Hook} {const char *} TARGET_ASM_CLOSE_PAREN
5934 These target hooks are C string constants, describing the syntax in the
5935 assembler for grouping arithmetic expressions. If not overridden, they
5936 default to normal parentheses, which is correct for most assemblers.
5939 These macros are provided by @file{real.h} for writing the definitions
5940 of @code{ASM_OUTPUT_DOUBLE} and the like:
5943 @item REAL_VALUE_TO_TARGET_SINGLE (@var{x}, @var{l})
5944 @itemx REAL_VALUE_TO_TARGET_DOUBLE (@var{x}, @var{l})
5945 @itemx REAL_VALUE_TO_TARGET_LONG_DOUBLE (@var{x}, @var{l})
5946 @findex REAL_VALUE_TO_TARGET_SINGLE
5947 @findex REAL_VALUE_TO_TARGET_DOUBLE
5948 @findex REAL_VALUE_TO_TARGET_LONG_DOUBLE
5949 These translate @var{x}, of type @code{REAL_VALUE_TYPE}, to the target's
5950 floating point representation, and store its bit pattern in the array of
5951 @code{long int} whose address is @var{l}. The number of elements in the
5952 output array is determined by the size of the desired target floating
5953 point data type: 32 bits of it go in each @code{long int} array
5954 element. Each array element holds 32 bits of the result, even if
5955 @code{long int} is wider than 32 bits on the host machine.
5957 The array element values are designed so that you can print them out
5958 using @code{fprintf} in the order they should appear in the target
5961 @item REAL_VALUE_TO_DECIMAL (@var{x}, @var{format}, @var{string})
5962 @findex REAL_VALUE_TO_DECIMAL
5963 This macro converts @var{x}, of type @code{REAL_VALUE_TYPE}, to a
5964 decimal number and stores it as a string into @var{string}.
5965 You must pass, as @var{string}, the address of a long enough block
5966 of space to hold the result.
5968 The argument @var{format} is a @code{printf}-specification that serves
5969 as a suggestion for how to format the output string.
5972 @node Uninitialized Data
5973 @subsection Output of Uninitialized Variables
5975 Each of the macros in this section is used to do the whole job of
5976 outputting a single uninitialized variable.
5979 @findex ASM_OUTPUT_COMMON
5980 @item ASM_OUTPUT_COMMON (@var{stream}, @var{name}, @var{size}, @var{rounded})
5981 A C statement (sans semicolon) to output to the stdio stream
5982 @var{stream} the assembler definition of a common-label named
5983 @var{name} whose size is @var{size} bytes. The variable @var{rounded}
5984 is the size rounded up to whatever alignment the caller wants.
5986 Use the expression @code{assemble_name (@var{stream}, @var{name})} to
5987 output the name itself; before and after that, output the additional
5988 assembler syntax for defining the name, and a newline.
5990 This macro controls how the assembler definitions of uninitialized
5991 common global variables are output.
5993 @findex ASM_OUTPUT_ALIGNED_COMMON
5994 @item ASM_OUTPUT_ALIGNED_COMMON (@var{stream}, @var{name}, @var{size}, @var{alignment})
5995 Like @code{ASM_OUTPUT_COMMON} except takes the required alignment as a
5996 separate, explicit argument. If you define this macro, it is used in
5997 place of @code{ASM_OUTPUT_COMMON}, and gives you more flexibility in
5998 handling the required alignment of the variable. The alignment is specified
5999 as the number of bits.
6001 @findex ASM_OUTPUT_ALIGNED_DECL_COMMON
6002 @item ASM_OUTPUT_ALIGNED_DECL_COMMON (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{alignment})
6003 Like @code{ASM_OUTPUT_ALIGNED_COMMON} except that @var{decl} of the
6004 variable to be output, if there is one, or @code{NULL_TREE} if there
6005 is no corresponding variable. If you define this macro, GCC will use it
6006 in place of both @code{ASM_OUTPUT_COMMON} and
6007 @code{ASM_OUTPUT_ALIGNED_COMMON}. Define this macro when you need to see
6008 the variable's decl in order to chose what to output.
6010 @findex ASM_OUTPUT_SHARED_COMMON
6011 @item ASM_OUTPUT_SHARED_COMMON (@var{stream}, @var{name}, @var{size}, @var{rounded})
6012 If defined, it is similar to @code{ASM_OUTPUT_COMMON}, except that it
6013 is used when @var{name} is shared. If not defined, @code{ASM_OUTPUT_COMMON}
6016 @findex ASM_OUTPUT_BSS
6017 @item ASM_OUTPUT_BSS (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{rounded})
6018 A C statement (sans semicolon) to output to the stdio stream
6019 @var{stream} the assembler definition of uninitialized global @var{decl} named
6020 @var{name} whose size is @var{size} bytes. The variable @var{rounded}
6021 is the size rounded up to whatever alignment the caller wants.
6023 Try to use function @code{asm_output_bss} defined in @file{varasm.c} when
6024 defining this macro. If unable, use the expression
6025 @code{assemble_name (@var{stream}, @var{name})} to output the name itself;
6026 before and after that, output the additional assembler syntax for defining
6027 the name, and a newline.
6029 This macro controls how the assembler definitions of uninitialized global
6030 variables are output. This macro exists to properly support languages like
6031 C++ which do not have @code{common} data. However, this macro currently
6032 is not defined for all targets. If this macro and
6033 @code{ASM_OUTPUT_ALIGNED_BSS} are not defined then @code{ASM_OUTPUT_COMMON}
6034 or @code{ASM_OUTPUT_ALIGNED_COMMON} or
6035 @code{ASM_OUTPUT_ALIGNED_DECL_COMMON} is used.
6037 @findex ASM_OUTPUT_ALIGNED_BSS
6038 @item ASM_OUTPUT_ALIGNED_BSS (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{alignment})
6039 Like @code{ASM_OUTPUT_BSS} except takes the required alignment as a
6040 separate, explicit argument. If you define this macro, it is used in
6041 place of @code{ASM_OUTPUT_BSS}, and gives you more flexibility in
6042 handling the required alignment of the variable. The alignment is specified
6043 as the number of bits.
6045 Try to use function @code{asm_output_aligned_bss} defined in file
6046 @file{varasm.c} when defining this macro.
6048 @findex ASM_OUTPUT_SHARED_BSS
6049 @item ASM_OUTPUT_SHARED_BSS (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{rounded})
6050 If defined, it is similar to @code{ASM_OUTPUT_BSS}, except that it
6051 is used when @var{name} is shared. If not defined, @code{ASM_OUTPUT_BSS}
6054 @findex ASM_OUTPUT_LOCAL
6055 @item ASM_OUTPUT_LOCAL (@var{stream}, @var{name}, @var{size}, @var{rounded})
6056 A C statement (sans semicolon) to output to the stdio stream
6057 @var{stream} the assembler definition of a local-common-label named
6058 @var{name} whose size is @var{size} bytes. The variable @var{rounded}
6059 is the size rounded up to whatever alignment the caller wants.
6061 Use the expression @code{assemble_name (@var{stream}, @var{name})} to
6062 output the name itself; before and after that, output the additional
6063 assembler syntax for defining the name, and a newline.
6065 This macro controls how the assembler definitions of uninitialized
6066 static variables are output.
6068 @findex ASM_OUTPUT_ALIGNED_LOCAL
6069 @item ASM_OUTPUT_ALIGNED_LOCAL (@var{stream}, @var{name}, @var{size}, @var{alignment})
6070 Like @code{ASM_OUTPUT_LOCAL} except takes the required alignment as a
6071 separate, explicit argument. If you define this macro, it is used in
6072 place of @code{ASM_OUTPUT_LOCAL}, and gives you more flexibility in
6073 handling the required alignment of the variable. The alignment is specified
6074 as the number of bits.
6076 @findex ASM_OUTPUT_ALIGNED_DECL_LOCAL
6077 @item ASM_OUTPUT_ALIGNED_DECL_LOCAL (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{alignment})
6078 Like @code{ASM_OUTPUT_ALIGNED_DECL} except that @var{decl} of the
6079 variable to be output, if there is one, or @code{NULL_TREE} if there
6080 is no corresponding variable. If you define this macro, GCC will use it
6081 in place of both @code{ASM_OUTPUT_DECL} and
6082 @code{ASM_OUTPUT_ALIGNED_DECL}. Define this macro when you need to see
6083 the variable's decl in order to chose what to output.
6085 @findex ASM_OUTPUT_SHARED_LOCAL
6086 @item ASM_OUTPUT_SHARED_LOCAL (@var{stream}, @var{name}, @var{size}, @var{rounded})
6087 If defined, it is similar to @code{ASM_OUTPUT_LOCAL}, except that it
6088 is used when @var{name} is shared. If not defined, @code{ASM_OUTPUT_LOCAL}
6093 @subsection Output and Generation of Labels
6095 @c prevent bad page break with this line
6096 This is about outputting labels.
6099 @findex ASM_OUTPUT_LABEL
6100 @findex assemble_name
6101 @item ASM_OUTPUT_LABEL (@var{stream}, @var{name})
6102 A C statement (sans semicolon) to output to the stdio stream
6103 @var{stream} the assembler definition of a label named @var{name}.
6104 Use the expression @code{assemble_name (@var{stream}, @var{name})} to
6105 output the name itself; before and after that, output the additional
6106 assembler syntax for defining the name, and a newline.
6108 @findex ASM_DECLARE_FUNCTION_NAME
6109 @item ASM_DECLARE_FUNCTION_NAME (@var{stream}, @var{name}, @var{decl})
6110 A C statement (sans semicolon) to output to the stdio stream
6111 @var{stream} any text necessary for declaring the name @var{name} of a
6112 function which is being defined. This macro is responsible for
6113 outputting the label definition (perhaps using
6114 @code{ASM_OUTPUT_LABEL}). The argument @var{decl} is the
6115 @code{FUNCTION_DECL} tree node representing the function.
6117 If this macro is not defined, then the function name is defined in the
6118 usual manner as a label (by means of @code{ASM_OUTPUT_LABEL}).
6120 @findex ASM_DECLARE_FUNCTION_SIZE
6121 @item ASM_DECLARE_FUNCTION_SIZE (@var{stream}, @var{name}, @var{decl})
6122 A C statement (sans semicolon) to output to the stdio stream
6123 @var{stream} any text necessary for declaring the size of a function
6124 which is being defined. The argument @var{name} is the name of the
6125 function. The argument @var{decl} is the @code{FUNCTION_DECL} tree node
6126 representing the function.
6128 If this macro is not defined, then the function size is not defined.
6130 @findex ASM_DECLARE_OBJECT_NAME
6131 @item ASM_DECLARE_OBJECT_NAME (@var{stream}, @var{name}, @var{decl})
6132 A C statement (sans semicolon) to output to the stdio stream
6133 @var{stream} any text necessary for declaring the name @var{name} of an
6134 initialized variable which is being defined. This macro must output the
6135 label definition (perhaps using @code{ASM_OUTPUT_LABEL}). The argument
6136 @var{decl} is the @code{VAR_DECL} tree node representing the variable.
6138 If this macro is not defined, then the variable name is defined in the
6139 usual manner as a label (by means of @code{ASM_OUTPUT_LABEL}).
6141 @findex ASM_DECLARE_REGISTER_GLOBAL
6142 @item ASM_DECLARE_REGISTER_GLOBAL (@var{stream}, @var{decl}, @var{regno}, @var{name})
6143 A C statement (sans semicolon) to output to the stdio stream
6144 @var{stream} any text necessary for claiming a register @var{regno}
6145 for a global variable @var{decl} with name @var{name}.
6147 If you don't define this macro, that is equivalent to defining it to do
6150 @findex ASM_FINISH_DECLARE_OBJECT
6151 @item ASM_FINISH_DECLARE_OBJECT (@var{stream}, @var{decl}, @var{toplevel}, @var{atend})
6152 A C statement (sans semicolon) to finish up declaring a variable name
6153 once the compiler has processed its initializer fully and thus has had a
6154 chance to determine the size of an array when controlled by an
6155 initializer. This is used on systems where it's necessary to declare
6156 something about the size of the object.
6158 If you don't define this macro, that is equivalent to defining it to do
6161 @findex ASM_GLOBALIZE_LABEL
6162 @item ASM_GLOBALIZE_LABEL (@var{stream}, @var{name})
6163 A C statement (sans semicolon) to output to the stdio stream
6164 @var{stream} some commands that will make the label @var{name} global;
6165 that is, available for reference from other files. Use the expression
6166 @code{assemble_name (@var{stream}, @var{name})} to output the name
6167 itself; before and after that, output the additional assembler syntax
6168 for making that name global, and a newline.
6170 @findex ASM_WEAKEN_LABEL
6171 @item ASM_WEAKEN_LABEL
6172 A C statement (sans semicolon) to output to the stdio stream
6173 @var{stream} some commands that will make the label @var{name} weak;
6174 that is, available for reference from other files but only used if
6175 no other definition is available. Use the expression
6176 @code{assemble_name (@var{stream}, @var{name})} to output the name
6177 itself; before and after that, output the additional assembler syntax
6178 for making that name weak, and a newline.
6180 If you don't define this macro, GCC will not support weak
6181 symbols and you should not define the @code{SUPPORTS_WEAK} macro.
6183 @findex SUPPORTS_WEAK
6185 A C expression which evaluates to true if the target supports weak symbols.
6187 If you don't define this macro, @file{defaults.h} provides a default
6188 definition. If @code{ASM_WEAKEN_LABEL} is defined, the default
6189 definition is @samp{1}; otherwise, it is @samp{0}. Define this macro if
6190 you want to control weak symbol support with a compiler flag such as
6193 @findex MAKE_DECL_ONE_ONLY (@var{decl})
6194 @item MAKE_DECL_ONE_ONLY
6195 A C statement (sans semicolon) to mark @var{decl} to be emitted as a
6196 public symbol such that extra copies in multiple translation units will
6197 be discarded by the linker. Define this macro if your object file
6198 format provides support for this concept, such as the @samp{COMDAT}
6199 section flags in the Microsoft Windows PE/COFF format, and this support
6200 requires changes to @var{decl}, such as putting it in a separate section.
6202 @findex SUPPORTS_ONE_ONLY
6203 @item SUPPORTS_ONE_ONLY
6204 A C expression which evaluates to true if the target supports one-only
6207 If you don't define this macro, @file{varasm.c} provides a default
6208 definition. If @code{MAKE_DECL_ONE_ONLY} is defined, the default
6209 definition is @samp{1}; otherwise, it is @samp{0}. Define this macro if
6210 you want to control one-only symbol support with a compiler flag, or if
6211 setting the @code{DECL_ONE_ONLY} flag is enough to mark a declaration to
6212 be emitted as one-only.
6214 @findex ASM_OUTPUT_EXTERNAL
6215 @item ASM_OUTPUT_EXTERNAL (@var{stream}, @var{decl}, @var{name})
6216 A C statement (sans semicolon) to output to the stdio stream
6217 @var{stream} any text necessary for declaring the name of an external
6218 symbol named @var{name} which is referenced in this compilation but
6219 not defined. The value of @var{decl} is the tree node for the
6222 This macro need not be defined if it does not need to output anything.
6223 The GNU assembler and most Unix assemblers don't require anything.
6225 @findex ASM_OUTPUT_EXTERNAL_LIBCALL
6226 @item ASM_OUTPUT_EXTERNAL_LIBCALL (@var{stream}, @var{symref})
6227 A C statement (sans semicolon) to output on @var{stream} an assembler
6228 pseudo-op to declare a library function name external. The name of the
6229 library function is given by @var{symref}, which has type @code{rtx} and
6230 is a @code{symbol_ref}.
6232 This macro need not be defined if it does not need to output anything.
6233 The GNU assembler and most Unix assemblers don't require anything.
6235 @findex ASM_OUTPUT_LABELREF
6236 @item ASM_OUTPUT_LABELREF (@var{stream}, @var{name})
6237 A C statement (sans semicolon) to output to the stdio stream
6238 @var{stream} a reference in assembler syntax to a label named
6239 @var{name}. This should add @samp{_} to the front of the name, if that
6240 is customary on your operating system, as it is in most Berkeley Unix
6241 systems. This macro is used in @code{assemble_name}.
6243 @findex ASM_OUTPUT_SYMBOL_REF
6244 @item ASM_OUTPUT_SYMBOL_REF (@var{stream}, @var{sym})
6245 A C statement (sans semicolon) to output a reference to
6246 @code{SYMBOL_REF} @var{sym}. If not defined, @code{assemble_name}
6247 will be used to output the name of the symbol. This macro may be used
6248 to modify the way a symbol is referenced depending on information
6249 encoded by @code{ENCODE_SECTION_INFO}.
6251 @findex ASM_OUTPUT_LABEL_REF
6252 @item ASM_OUTPUT_LABEL_REF (@var{stream}, @var{buf})
6253 A C statement (sans semicolon) to output a reference to @var{buf}, the
6254 result of ASM_GENERATE_INTERNAL_LABEL. If not defined,
6255 @code{assemble_name} will be used to output the name of the symbol.
6256 This macro is not used by @code{output_asm_label}, or the @code{%l}
6257 specifier that calls it; the intention is that this macro should be set
6258 when it is necessary to output a label differently when its address
6261 @findex ASM_OUTPUT_INTERNAL_LABEL
6262 @item ASM_OUTPUT_INTERNAL_LABEL (@var{stream}, @var{prefix}, @var{num})
6263 A C statement to output to the stdio stream @var{stream} a label whose
6264 name is made from the string @var{prefix} and the number @var{num}.
6266 It is absolutely essential that these labels be distinct from the labels
6267 used for user-level functions and variables. Otherwise, certain programs
6268 will have name conflicts with internal labels.
6270 It is desirable to exclude internal labels from the symbol table of the
6271 object file. Most assemblers have a naming convention for labels that
6272 should be excluded; on many systems, the letter @samp{L} at the
6273 beginning of a label has this effect. You should find out what
6274 convention your system uses, and follow it.
6276 The usual definition of this macro is as follows:
6279 fprintf (@var{stream}, "L%s%d:\n", @var{prefix}, @var{num})
6282 @findex ASM_OUTPUT_DEBUG_LABEL
6283 @item ASM_OUTPUT_DEBUG_LABEL (@var{stream}, @var{prefix}, @var{num})
6284 A C statement to output to the stdio stream @var{stream} a debug info
6285 label whose name is made from the string @var{prefix} and the number
6286 @var{num}. This is useful for VLIW targets, where debug info labels
6287 may need to be treated differently than branch target labels. On some
6288 systems, branch target labels must be at the beginning of instruction
6289 bundles, but debug info labels can occur in the middle of instruction
6292 If this macro is not defined, then @code{ASM_OUTPUT_INTERNAL_LABEL} will be
6295 @findex ASM_OUTPUT_ALTERNATE_LABEL_NAME
6296 @item ASM_OUTPUT_ALTERNATE_LABEL_NAME (@var{stream}, @var{string})
6297 A C statement to output to the stdio stream @var{stream} the string
6300 The default definition of this macro is as follows:
6303 fprintf (@var{stream}, "%s:\n", LABEL_ALTERNATE_NAME (INSN))
6306 @findex ASM_GENERATE_INTERNAL_LABEL
6307 @item ASM_GENERATE_INTERNAL_LABEL (@var{string}, @var{prefix}, @var{num})
6308 A C statement to store into the string @var{string} a label whose name
6309 is made from the string @var{prefix} and the number @var{num}.
6311 This string, when output subsequently by @code{assemble_name}, should
6312 produce the output that @code{ASM_OUTPUT_INTERNAL_LABEL} would produce
6313 with the same @var{prefix} and @var{num}.
6315 If the string begins with @samp{*}, then @code{assemble_name} will
6316 output the rest of the string unchanged. It is often convenient for
6317 @code{ASM_GENERATE_INTERNAL_LABEL} to use @samp{*} in this way. If the
6318 string doesn't start with @samp{*}, then @code{ASM_OUTPUT_LABELREF} gets
6319 to output the string, and may change it. (Of course,
6320 @code{ASM_OUTPUT_LABELREF} is also part of your machine description, so
6321 you should know what it does on your machine.)
6323 @findex ASM_FORMAT_PRIVATE_NAME
6324 @item ASM_FORMAT_PRIVATE_NAME (@var{outvar}, @var{name}, @var{number})
6325 A C expression to assign to @var{outvar} (which is a variable of type
6326 @code{char *}) a newly allocated string made from the string
6327 @var{name} and the number @var{number}, with some suitable punctuation
6328 added. Use @code{alloca} to get space for the string.
6330 The string will be used as an argument to @code{ASM_OUTPUT_LABELREF} to
6331 produce an assembler label for an internal static variable whose name is
6332 @var{name}. Therefore, the string must be such as to result in valid
6333 assembler code. The argument @var{number} is different each time this
6334 macro is executed; it prevents conflicts between similarly-named
6335 internal static variables in different scopes.
6337 Ideally this string should not be a valid C identifier, to prevent any
6338 conflict with the user's own symbols. Most assemblers allow periods
6339 or percent signs in assembler symbols; putting at least one of these
6340 between the name and the number will suffice.
6342 @findex ASM_OUTPUT_DEF
6343 @item ASM_OUTPUT_DEF (@var{stream}, @var{name}, @var{value})
6344 A C statement to output to the stdio stream @var{stream} assembler code
6345 which defines (equates) the symbol @var{name} to have the value @var{value}.
6348 If @code{SET_ASM_OP} is defined, a default definition is provided which is
6349 correct for most systems.
6351 @findex ASM_OUTPUT_DEF_FROM_DECLS
6352 @item ASM_OUTPUT_DEF_FROM_DECLS (@var{stream}, @var{decl_of_name}, @var{decl_of_value})
6353 A C statement to output to the stdio stream @var{stream} assembler code
6354 which defines (equates) the symbol whose tree node is @var{decl_of_name}
6355 to have the value of the tree node @var{decl_of_value}. This macro will
6356 be used in preference to @samp{ASM_OUTPUT_DEF} if it is defined and if
6357 the tree nodes are available.
6359 @findex ASM_OUTPUT_DEFINE_LABEL_DIFFERENCE_SYMBOL
6360 @item ASM_OUTPUT_DEFINE_LABEL_DIFFERENCE_SYMBOL (@var{stream}, @var{symbol}, @var{high}, @var{low})
6361 A C statement to output to the stdio stream @var{stream} assembler code
6362 which defines (equates) the symbol @var{symbol} to have a value equal to
6363 the difference of the two symbols @var{high} and @var{low},
6364 i.e.@: @var{high} minus @var{low}. GCC guarantees that the symbols @var{high}
6365 and @var{low} are already known by the assembler so that the difference
6366 resolves into a constant.
6369 If @code{SET_ASM_OP} is defined, a default definition is provided which is
6370 correct for most systems.
6372 @findex ASM_OUTPUT_WEAK_ALIAS
6373 @item ASM_OUTPUT_WEAK_ALIAS (@var{stream}, @var{name}, @var{value})
6374 A C statement to output to the stdio stream @var{stream} assembler code
6375 which defines (equates) the weak symbol @var{name} to have the value
6376 @var{value}. If @var{value} is @code{NULL}, it defines @var{name} as
6377 an undefined weak symbol.
6379 Define this macro if the target only supports weak aliases; define
6380 @code{ASM_OUTPUT_DEF} instead if possible.
6382 @findex OBJC_GEN_METHOD_LABEL
6383 @item OBJC_GEN_METHOD_LABEL (@var{buf}, @var{is_inst}, @var{class_name}, @var{cat_name}, @var{sel_name})
6384 Define this macro to override the default assembler names used for
6385 Objective-C methods.
6387 The default name is a unique method number followed by the name of the
6388 class (e.g.@: @samp{_1_Foo}). For methods in categories, the name of
6389 the category is also included in the assembler name (e.g.@:
6392 These names are safe on most systems, but make debugging difficult since
6393 the method's selector is not present in the name. Therefore, particular
6394 systems define other ways of computing names.
6396 @var{buf} is an expression of type @code{char *} which gives you a
6397 buffer in which to store the name; its length is as long as
6398 @var{class_name}, @var{cat_name} and @var{sel_name} put together, plus
6399 50 characters extra.
6401 The argument @var{is_inst} specifies whether the method is an instance
6402 method or a class method; @var{class_name} is the name of the class;
6403 @var{cat_name} is the name of the category (or @code{NULL} if the method is not
6404 in a category); and @var{sel_name} is the name of the selector.
6406 On systems where the assembler can handle quoted names, you can use this
6407 macro to provide more human-readable names.
6409 @findex ASM_DECLARE_CLASS_REFERENCE
6410 @item ASM_DECLARE_CLASS_REFERENCE (@var{stream}, @var{name})
6411 A C statement (sans semicolon) to output to the stdio stream
6412 @var{stream} commands to declare that the label @var{name} is an
6413 Objective-C class reference. This is only needed for targets whose
6414 linkers have special support for NeXT-style runtimes.
6416 @findex ASM_DECLARE_UNRESOLVED_REFERENCE
6417 @item ASM_DECLARE_UNRESOLVED_REFERENCE (@var{stream}, @var{name})
6418 A C statement (sans semicolon) to output to the stdio stream
6419 @var{stream} commands to declare that the label @var{name} is an
6420 unresolved Objective-C class reference. This is only needed for targets
6421 whose linkers have special support for NeXT-style runtimes.
6424 @node Initialization
6425 @subsection How Initialization Functions Are Handled
6426 @cindex initialization routines
6427 @cindex termination routines
6428 @cindex constructors, output of
6429 @cindex destructors, output of
6431 The compiled code for certain languages includes @dfn{constructors}
6432 (also called @dfn{initialization routines})---functions to initialize
6433 data in the program when the program is started. These functions need
6434 to be called before the program is ``started''---that is to say, before
6435 @code{main} is called.
6437 Compiling some languages generates @dfn{destructors} (also called
6438 @dfn{termination routines}) that should be called when the program
6441 To make the initialization and termination functions work, the compiler
6442 must output something in the assembler code to cause those functions to
6443 be called at the appropriate time. When you port the compiler to a new
6444 system, you need to specify how to do this.
6446 There are two major ways that GCC currently supports the execution of
6447 initialization and termination functions. Each way has two variants.
6448 Much of the structure is common to all four variations.
6450 @findex __CTOR_LIST__
6451 @findex __DTOR_LIST__
6452 The linker must build two lists of these functions---a list of
6453 initialization functions, called @code{__CTOR_LIST__}, and a list of
6454 termination functions, called @code{__DTOR_LIST__}.
6456 Each list always begins with an ignored function pointer (which may hold
6457 0, @minus{}1, or a count of the function pointers after it, depending on
6458 the environment). This is followed by a series of zero or more function
6459 pointers to constructors (or destructors), followed by a function
6460 pointer containing zero.
6462 Depending on the operating system and its executable file format, either
6463 @file{crtstuff.c} or @file{libgcc2.c} traverses these lists at startup
6464 time and exit time. Constructors are called in reverse order of the
6465 list; destructors in forward order.
6467 The best way to handle static constructors works only for object file
6468 formats which provide arbitrarily-named sections. A section is set
6469 aside for a list of constructors, and another for a list of destructors.
6470 Traditionally these are called @samp{.ctors} and @samp{.dtors}. Each
6471 object file that defines an initialization function also puts a word in
6472 the constructor section to point to that function. The linker
6473 accumulates all these words into one contiguous @samp{.ctors} section.
6474 Termination functions are handled similarly.
6476 This method will be chosen as the default by @file{target-def.h} if
6477 @code{TARGET_ASM_NAMED_SECTION} is defined. A target that does not
6478 support arbitrary sections, but does support special designated
6479 constructor and destructor sections may define @code{CTORS_SECTION_ASM_OP}
6480 and @code{DTORS_SECTION_ASM_OP} to achieve the same effect.
6482 When arbitrary sections are available, there are two variants, depending
6483 upon how the code in @file{crtstuff.c} is called. On systems that
6484 support a @dfn{.init} section which is executed at program startup,
6485 parts of @file{crtstuff.c} are compiled into that section. The
6486 program is linked by the @code{gcc} driver like this:
6489 ld -o @var{output_file} crti.o crtbegin.o @dots{} -lgcc crtend.o crtn.o
6492 The prologue of a function (@code{__init}) appears in the @code{.init}
6493 section of @file{crti.o}; the epilogue appears in @file{crtn.o}. Likewise
6494 for the function @code{__fini} in the @dfn{.fini} section. Normally these
6495 files are provided by the operating system or by the GNU C library, but
6496 are provided by GCC for a few targets.
6498 The objects @file{crtbegin.o} and @file{crtend.o} are (for most targets)
6499 compiled from @file{crtstuff.c}. They contain, among other things, code
6500 fragments within the @code{.init} and @code{.fini} sections that branch
6501 to routines in the @code{.text} section. The linker will pull all parts
6502 of a section together, which results in a complete @code{__init} function
6503 that invokes the routines we need at startup.
6505 To use this variant, you must define the @code{INIT_SECTION_ASM_OP}
6508 If no init section is available, when GCC compiles any function called
6509 @code{main} (or more accurately, any function designated as a program
6510 entry point by the language front end calling @code{expand_main_function}),
6511 it inserts a procedure call to @code{__main} as the first executable code
6512 after the function prologue. The @code{__main} function is defined
6513 in @file{libgcc2.c} and runs the global constructors.
6515 In file formats that don't support arbitrary sections, there are again
6516 two variants. In the simplest variant, the GNU linker (GNU @code{ld})
6517 and an `a.out' format must be used. In this case,
6518 @code{TARGET_ASM_CONSTRUCTOR} is defined to produce a @code{.stabs}
6519 entry of type @samp{N_SETT}, referencing the name @code{__CTOR_LIST__},
6520 and with the address of the void function containing the initialization
6521 code as its value. The GNU linker recognizes this as a request to add
6522 the value to a @dfn{set}; the values are accumulated, and are eventually
6523 placed in the executable as a vector in the format described above, with
6524 a leading (ignored) count and a trailing zero element.
6525 @code{TARGET_ASM_DESTRUCTOR} is handled similarly. Since no init
6526 section is available, the absence of @code{INIT_SECTION_ASM_OP} causes
6527 the compilation of @code{main} to call @code{__main} as above, starting
6528 the initialization process.
6530 The last variant uses neither arbitrary sections nor the GNU linker.
6531 This is preferable when you want to do dynamic linking and when using
6532 file formats which the GNU linker does not support, such as `ECOFF'@. In
6533 this case, @code{TARGET_HAVE_CTORS_DTORS} is false, initialization and
6534 termination functions are recognized simply by their names. This requires
6535 an extra program in the linkage step, called @command{collect2}. This program
6536 pretends to be the linker, for use with GCC; it does its job by running
6537 the ordinary linker, but also arranges to include the vectors of
6538 initialization and termination functions. These functions are called
6539 via @code{__main} as described above. In order to use this method,
6540 @code{use_collect2} must be defined in the target in @file{config.gcc}.
6543 The following section describes the specific macros that control and
6544 customize the handling of initialization and termination functions.
6547 @node Macros for Initialization
6548 @subsection Macros Controlling Initialization Routines
6550 Here are the macros that control how the compiler handles initialization
6551 and termination functions:
6554 @findex INIT_SECTION_ASM_OP
6555 @item INIT_SECTION_ASM_OP
6556 If defined, a C string constant, including spacing, for the assembler
6557 operation to identify the following data as initialization code. If not
6558 defined, GCC will assume such a section does not exist. When you are
6559 using special sections for initialization and termination functions, this
6560 macro also controls how @file{crtstuff.c} and @file{libgcc2.c} arrange to
6561 run the initialization functions.
6563 @item HAS_INIT_SECTION
6564 @findex HAS_INIT_SECTION
6565 If defined, @code{main} will not call @code{__main} as described above.
6566 This macro should be defined for systems that control start-up code
6567 on a symbol-by-symbol basis, such as OSF/1, and should not
6568 be defined explicitly for systems that support @code{INIT_SECTION_ASM_OP}.
6570 @item LD_INIT_SWITCH
6571 @findex LD_INIT_SWITCH
6572 If defined, a C string constant for a switch that tells the linker that
6573 the following symbol is an initialization routine.
6575 @item LD_FINI_SWITCH
6576 @findex LD_FINI_SWITCH
6577 If defined, a C string constant for a switch that tells the linker that
6578 the following symbol is a finalization routine.
6580 @item COLLECT_SHARED_INIT_FUNC (@var{stream}, @var{func})
6581 If defined, a C statement that will write a function that can be
6582 automatically called when a shared library is loaded. The function
6583 should call @var{func}, which takes no arguments. If not defined, and
6584 the object format requires an explicit initialization function, then a
6585 function called @code{_GLOBAL__DI} will be generated.
6587 This function and the following one are used by collect2 when linking a
6588 shared library that needs constructors or destructors, or has DWARF2
6589 exception tables embedded in the code.
6591 @item COLLECT_SHARED_FINI_FUNC (@var{stream}, @var{func})
6592 If defined, a C statement that will write a function that can be
6593 automatically called when a shared library is unloaded. The function
6594 should call @var{func}, which takes no arguments. If not defined, and
6595 the object format requires an explicit finalization function, then a
6596 function called @code{_GLOBAL__DD} will be generated.
6599 @findex INVOKE__main
6600 If defined, @code{main} will call @code{__main} despite the presence of
6601 @code{INIT_SECTION_ASM_OP}. This macro should be defined for systems
6602 where the init section is not actually run automatically, but is still
6603 useful for collecting the lists of constructors and destructors.
6605 @item SUPPORTS_INIT_PRIORITY
6606 @findex SUPPORTS_INIT_PRIORITY
6607 If nonzero, the C++ @code{init_priority} attribute is supported and the
6608 compiler should emit instructions to control the order of initialization
6609 of objects. If zero, the compiler will issue an error message upon
6610 encountering an @code{init_priority} attribute.
6613 @deftypefn {Target Hook} bool TARGET_HAVE_CTORS_DTORS
6614 This value is true if the target supports some ``native'' method of
6615 collecting constructors and destructors to be run at startup and exit.
6616 It is false if we must use @command{collect2}.
6619 @deftypefn {Target Hook} void TARGET_ASM_CONSTRUCTOR (rtx @var{symbol}, int @var{priority})
6620 If defined, a function that outputs assembler code to arrange to call
6621 the function referenced by @var{symbol} at initialization time.
6623 Assume that @var{symbol} is a @code{SYMBOL_REF} for a function taking
6624 no arguments and with no return value. If the target supports initialization
6625 priorities, @var{priority} is a value between 0 and @code{MAX_INIT_PRIORITY};
6626 otherwise it must be @code{DEFAULT_INIT_PRIORITY}.
6628 If this macro is not defined by the target, a suitable default will
6629 be chosen if (1) the target supports arbitrary section names, (2) the
6630 target defines @code{CTORS_SECTION_ASM_OP}, or (3) @code{USE_COLLECT2}
6634 @deftypefn {Target Hook} void TARGET_ASM_DESTRUCTOR (rtx @var{symbol}, int @var{priority})
6635 This is like @code{TARGET_ASM_CONSTRUCTOR} but used for termination
6636 functions rather than initialization functions.
6639 If @code{TARGET_HAVE_CTORS_DTORS} is true, the initialization routine
6640 generated for the generated object file will have static linkage.
6642 If your system uses @command{collect2} as the means of processing
6643 constructors, then that program normally uses @command{nm} to scan
6644 an object file for constructor functions to be called.
6646 On certain kinds of systems, you can define these macros to make
6647 @command{collect2} work faster (and, in some cases, make it work at all):
6650 @findex OBJECT_FORMAT_COFF
6651 @item OBJECT_FORMAT_COFF
6652 Define this macro if the system uses COFF (Common Object File Format)
6653 object files, so that @command{collect2} can assume this format and scan
6654 object files directly for dynamic constructor/destructor functions.
6656 @findex OBJECT_FORMAT_ROSE
6657 @item OBJECT_FORMAT_ROSE
6658 Define this macro if the system uses ROSE format object files, so that
6659 @command{collect2} can assume this format and scan object files directly
6660 for dynamic constructor/destructor functions.
6662 These macros are effective only in a native compiler; @command{collect2} as
6663 part of a cross compiler always uses @command{nm} for the target machine.
6665 @findex REAL_NM_FILE_NAME
6666 @item REAL_NM_FILE_NAME
6667 Define this macro as a C string constant containing the file name to use
6668 to execute @command{nm}. The default is to search the path normally for
6671 If your system supports shared libraries and has a program to list the
6672 dynamic dependencies of a given library or executable, you can define
6673 these macros to enable support for running initialization and
6674 termination functions in shared libraries:
6678 Define this macro to a C string constant containing the name of the program
6679 which lists dynamic dependencies, like @command{"ldd"} under SunOS 4.
6681 @findex PARSE_LDD_OUTPUT
6682 @item PARSE_LDD_OUTPUT (@var{ptr})
6683 Define this macro to be C code that extracts filenames from the output
6684 of the program denoted by @code{LDD_SUFFIX}. @var{ptr} is a variable
6685 of type @code{char *} that points to the beginning of a line of output
6686 from @code{LDD_SUFFIX}. If the line lists a dynamic dependency, the
6687 code must advance @var{ptr} to the beginning of the filename on that
6688 line. Otherwise, it must set @var{ptr} to @code{NULL}.
6691 @node Instruction Output
6692 @subsection Output of Assembler Instructions
6694 @c prevent bad page break with this line
6695 This describes assembler instruction output.
6698 @findex REGISTER_NAMES
6699 @item REGISTER_NAMES
6700 A C initializer containing the assembler's names for the machine
6701 registers, each one as a C string constant. This is what translates
6702 register numbers in the compiler into assembler language.
6704 @findex ADDITIONAL_REGISTER_NAMES
6705 @item ADDITIONAL_REGISTER_NAMES
6706 If defined, a C initializer for an array of structures containing a name
6707 and a register number. This macro defines additional names for hard
6708 registers, thus allowing the @code{asm} option in declarations to refer
6709 to registers using alternate names.
6711 @findex ASM_OUTPUT_OPCODE
6712 @item ASM_OUTPUT_OPCODE (@var{stream}, @var{ptr})
6713 Define this macro if you are using an unusual assembler that
6714 requires different names for the machine instructions.
6716 The definition is a C statement or statements which output an
6717 assembler instruction opcode to the stdio stream @var{stream}. The
6718 macro-operand @var{ptr} is a variable of type @code{char *} which
6719 points to the opcode name in its ``internal'' form---the form that is
6720 written in the machine description. The definition should output the
6721 opcode name to @var{stream}, performing any translation you desire, and
6722 increment the variable @var{ptr} to point at the end of the opcode
6723 so that it will not be output twice.
6725 In fact, your macro definition may process less than the entire opcode
6726 name, or more than the opcode name; but if you want to process text
6727 that includes @samp{%}-sequences to substitute operands, you must take
6728 care of the substitution yourself. Just be sure to increment
6729 @var{ptr} over whatever text should not be output normally.
6731 @findex recog_data.operand
6732 If you need to look at the operand values, they can be found as the
6733 elements of @code{recog_data.operand}.
6735 If the macro definition does nothing, the instruction is output
6738 @findex FINAL_PRESCAN_INSN
6739 @item FINAL_PRESCAN_INSN (@var{insn}, @var{opvec}, @var{noperands})
6740 If defined, a C statement to be executed just prior to the output of
6741 assembler code for @var{insn}, to modify the extracted operands so
6742 they will be output differently.
6744 Here the argument @var{opvec} is the vector containing the operands
6745 extracted from @var{insn}, and @var{noperands} is the number of
6746 elements of the vector which contain meaningful data for this insn.
6747 The contents of this vector are what will be used to convert the insn
6748 template into assembler code, so you can change the assembler output
6749 by changing the contents of the vector.
6751 This macro is useful when various assembler syntaxes share a single
6752 file of instruction patterns; by defining this macro differently, you
6753 can cause a large class of instructions to be output differently (such
6754 as with rearranged operands). Naturally, variations in assembler
6755 syntax affecting individual insn patterns ought to be handled by
6756 writing conditional output routines in those patterns.
6758 If this macro is not defined, it is equivalent to a null statement.
6760 @findex FINAL_PRESCAN_LABEL
6761 @item FINAL_PRESCAN_LABEL
6762 If defined, @code{FINAL_PRESCAN_INSN} will be called on each
6763 @code{CODE_LABEL}. In that case, @var{opvec} will be a null pointer and
6764 @var{noperands} will be zero.
6766 @findex PRINT_OPERAND
6767 @item PRINT_OPERAND (@var{stream}, @var{x}, @var{code})
6768 A C compound statement to output to stdio stream @var{stream} the
6769 assembler syntax for an instruction operand @var{x}. @var{x} is an
6772 @var{code} is a value that can be used to specify one of several ways
6773 of printing the operand. It is used when identical operands must be
6774 printed differently depending on the context. @var{code} comes from
6775 the @samp{%} specification that was used to request printing of the
6776 operand. If the specification was just @samp{%@var{digit}} then
6777 @var{code} is 0; if the specification was @samp{%@var{ltr}
6778 @var{digit}} then @var{code} is the ASCII code for @var{ltr}.
6781 If @var{x} is a register, this macro should print the register's name.
6782 The names can be found in an array @code{reg_names} whose type is
6783 @code{char *[]}. @code{reg_names} is initialized from
6784 @code{REGISTER_NAMES}.
6786 When the machine description has a specification @samp{%@var{punct}}
6787 (a @samp{%} followed by a punctuation character), this macro is called
6788 with a null pointer for @var{x} and the punctuation character for
6791 @findex PRINT_OPERAND_PUNCT_VALID_P
6792 @item PRINT_OPERAND_PUNCT_VALID_P (@var{code})
6793 A C expression which evaluates to true if @var{code} is a valid
6794 punctuation character for use in the @code{PRINT_OPERAND} macro. If
6795 @code{PRINT_OPERAND_PUNCT_VALID_P} is not defined, it means that no
6796 punctuation characters (except for the standard one, @samp{%}) are used
6799 @findex PRINT_OPERAND_ADDRESS
6800 @item PRINT_OPERAND_ADDRESS (@var{stream}, @var{x})
6801 A C compound statement to output to stdio stream @var{stream} the
6802 assembler syntax for an instruction operand that is a memory reference
6803 whose address is @var{x}. @var{x} is an RTL expression.
6805 @cindex @code{ENCODE_SECTION_INFO} usage
6806 On some machines, the syntax for a symbolic address depends on the
6807 section that the address refers to. On these machines, define the macro
6808 @code{ENCODE_SECTION_INFO} to store the information into the
6809 @code{symbol_ref}, and then check for it here. @xref{Assembler Format}.
6811 @findex DBR_OUTPUT_SEQEND
6812 @findex dbr_sequence_length
6813 @item DBR_OUTPUT_SEQEND(@var{file})
6814 A C statement, to be executed after all slot-filler instructions have
6815 been output. If necessary, call @code{dbr_sequence_length} to
6816 determine the number of slots filled in a sequence (zero if not
6817 currently outputting a sequence), to decide how many no-ops to output,
6820 Don't define this macro if it has nothing to do, but it is helpful in
6821 reading assembly output if the extent of the delay sequence is made
6822 explicit (e.g.@: with white space).
6824 @findex final_sequence
6825 Note that output routines for instructions with delay slots must be
6826 prepared to deal with not being output as part of a sequence
6827 (i.e.@: when the scheduling pass is not run, or when no slot fillers could be
6828 found.) The variable @code{final_sequence} is null when not
6829 processing a sequence, otherwise it contains the @code{sequence} rtx
6832 @findex REGISTER_PREFIX
6833 @findex LOCAL_LABEL_PREFIX
6834 @findex USER_LABEL_PREFIX
6835 @findex IMMEDIATE_PREFIX
6837 @item REGISTER_PREFIX
6838 @itemx LOCAL_LABEL_PREFIX
6839 @itemx USER_LABEL_PREFIX
6840 @itemx IMMEDIATE_PREFIX
6841 If defined, C string expressions to be used for the @samp{%R}, @samp{%L},
6842 @samp{%U}, and @samp{%I} options of @code{asm_fprintf} (see
6843 @file{final.c}). These are useful when a single @file{md} file must
6844 support multiple assembler formats. In that case, the various @file{tm.h}
6845 files can define these macros differently.
6847 @item ASM_FPRINTF_EXTENSIONS(@var{file}, @var{argptr}, @var{format})
6848 @findex ASM_FPRINTF_EXTENSIONS
6849 If defined this macro should expand to a series of @code{case}
6850 statements which will be parsed inside the @code{switch} statement of
6851 the @code{asm_fprintf} function. This allows targets to define extra
6852 printf formats which may useful when generating their assembler
6853 statements. Note that upper case letters are reserved for future
6854 generic extensions to asm_fprintf, and so are not available to target
6855 specific code. The output file is given by the parameter @var{file}.
6856 The varargs input pointer is @var{argptr} and the rest of the format
6857 string, starting the character after the one that is being switched
6858 upon, is pointed to by @var{format}.
6860 @findex ASSEMBLER_DIALECT
6861 @item ASSEMBLER_DIALECT
6862 If your target supports multiple dialects of assembler language (such as
6863 different opcodes), define this macro as a C expression that gives the
6864 numeric index of the assembler language dialect to use, with zero as the
6867 If this macro is defined, you may use constructs of the form
6869 @samp{@{option0|option1|option2@dots{}@}}
6872 in the output templates of patterns (@pxref{Output Template}) or in the
6873 first argument of @code{asm_fprintf}. This construct outputs
6874 @samp{option0}, @samp{option1}, @samp{option2}, etc., if the value of
6875 @code{ASSEMBLER_DIALECT} is zero, one, two, etc. Any special characters
6876 within these strings retain their usual meaning. If there are fewer
6877 alternatives within the braces than the value of
6878 @code{ASSEMBLER_DIALECT}, the construct outputs nothing.
6880 If you do not define this macro, the characters @samp{@{}, @samp{|} and
6881 @samp{@}} do not have any special meaning when used in templates or
6882 operands to @code{asm_fprintf}.
6884 Define the macros @code{REGISTER_PREFIX}, @code{LOCAL_LABEL_PREFIX},
6885 @code{USER_LABEL_PREFIX} and @code{IMMEDIATE_PREFIX} if you can express
6886 the variations in assembler language syntax with that mechanism. Define
6887 @code{ASSEMBLER_DIALECT} and use the @samp{@{option0|option1@}} syntax
6888 if the syntax variant are larger and involve such things as different
6889 opcodes or operand order.
6891 @findex ASM_OUTPUT_REG_PUSH
6892 @item ASM_OUTPUT_REG_PUSH (@var{stream}, @var{regno})
6893 A C expression to output to @var{stream} some assembler code
6894 which will push hard register number @var{regno} onto the stack.
6895 The code need not be optimal, since this macro is used only when
6898 @findex ASM_OUTPUT_REG_POP
6899 @item ASM_OUTPUT_REG_POP (@var{stream}, @var{regno})
6900 A C expression to output to @var{stream} some assembler code
6901 which will pop hard register number @var{regno} off of the stack.
6902 The code need not be optimal, since this macro is used only when
6906 @node Dispatch Tables
6907 @subsection Output of Dispatch Tables
6909 @c prevent bad page break with this line
6910 This concerns dispatch tables.
6913 @cindex dispatch table
6914 @findex ASM_OUTPUT_ADDR_DIFF_ELT
6915 @item ASM_OUTPUT_ADDR_DIFF_ELT (@var{stream}, @var{body}, @var{value}, @var{rel})
6916 A C statement to output to the stdio stream @var{stream} an assembler
6917 pseudo-instruction to generate a difference between two labels.
6918 @var{value} and @var{rel} are the numbers of two internal labels. The
6919 definitions of these labels are output using
6920 @code{ASM_OUTPUT_INTERNAL_LABEL}, and they must be printed in the same
6921 way here. For example,
6924 fprintf (@var{stream}, "\t.word L%d-L%d\n",
6925 @var{value}, @var{rel})
6928 You must provide this macro on machines where the addresses in a
6929 dispatch table are relative to the table's own address. If defined, GCC
6930 will also use this macro on all machines when producing PIC@.
6931 @var{body} is the body of the @code{ADDR_DIFF_VEC}; it is provided so that the
6932 mode and flags can be read.
6934 @findex ASM_OUTPUT_ADDR_VEC_ELT
6935 @item ASM_OUTPUT_ADDR_VEC_ELT (@var{stream}, @var{value})
6936 This macro should be provided on machines where the addresses
6937 in a dispatch table are absolute.
6939 The definition should be a C statement to output to the stdio stream
6940 @var{stream} an assembler pseudo-instruction to generate a reference to
6941 a label. @var{value} is the number of an internal label whose
6942 definition is output using @code{ASM_OUTPUT_INTERNAL_LABEL}.
6946 fprintf (@var{stream}, "\t.word L%d\n", @var{value})
6949 @findex ASM_OUTPUT_CASE_LABEL
6950 @item ASM_OUTPUT_CASE_LABEL (@var{stream}, @var{prefix}, @var{num}, @var{table})
6951 Define this if the label before a jump-table needs to be output
6952 specially. The first three arguments are the same as for
6953 @code{ASM_OUTPUT_INTERNAL_LABEL}; the fourth argument is the
6954 jump-table which follows (a @code{jump_insn} containing an
6955 @code{addr_vec} or @code{addr_diff_vec}).
6957 This feature is used on system V to output a @code{swbeg} statement
6960 If this macro is not defined, these labels are output with
6961 @code{ASM_OUTPUT_INTERNAL_LABEL}.
6963 @findex ASM_OUTPUT_CASE_END
6964 @item ASM_OUTPUT_CASE_END (@var{stream}, @var{num}, @var{table})
6965 Define this if something special must be output at the end of a
6966 jump-table. The definition should be a C statement to be executed
6967 after the assembler code for the table is written. It should write
6968 the appropriate code to stdio stream @var{stream}. The argument
6969 @var{table} is the jump-table insn, and @var{num} is the label-number
6970 of the preceding label.
6972 If this macro is not defined, nothing special is output at the end of
6976 @node Exception Region Output
6977 @subsection Assembler Commands for Exception Regions
6979 @c prevent bad page break with this line
6981 This describes commands marking the start and the end of an exception
6985 @findex EH_FRAME_SECTION_NAME
6986 @item EH_FRAME_SECTION_NAME
6987 If defined, a C string constant for the name of the section containing
6988 exception handling frame unwind information. If not defined, GCC will
6989 provide a default definition if the target supports named sections.
6990 @file{crtstuff.c} uses this macro to switch to the appropriate section.
6992 You should define this symbol if your target supports DWARF 2 frame
6993 unwind information and the default definition does not work.
6995 @findex EH_FRAME_IN_DATA_SECTION
6996 @item EH_FRAME_IN_DATA_SECTION
6997 If defined, DWARF 2 frame unwind information will be placed in the
6998 data section even though the target supports named sections. This
6999 might be necessary, for instance, if the system linker does garbage
7000 collection and sections cannot be marked as not to be collected.
7002 Do not define this macro unless @code{TARGET_ASM_NAMED_SECTION} is
7005 @findex MASK_RETURN_ADDR
7006 @item MASK_RETURN_ADDR
7007 An rtx used to mask the return address found via @code{RETURN_ADDR_RTX}, so
7008 that it does not contain any extraneous set bits in it.
7010 @findex DWARF2_UNWIND_INFO
7011 @item DWARF2_UNWIND_INFO
7012 Define this macro to 0 if your target supports DWARF 2 frame unwind
7013 information, but it does not yet work with exception handling.
7014 Otherwise, if your target supports this information (if it defines
7015 @samp{INCOMING_RETURN_ADDR_RTX} and either @samp{UNALIGNED_INT_ASM_OP}
7016 or @samp{OBJECT_FORMAT_ELF}), GCC will provide a default definition of
7019 If this macro is defined to 1, the DWARF 2 unwinder will be the default
7020 exception handling mechanism; otherwise, @code{setjmp}/@code{longjmp} will be used by
7023 If this macro is defined to anything, the DWARF 2 unwinder will be used
7024 instead of inline unwinders and @code{__unwind_function} in the non-@code{setjmp} case.
7026 @findex DWARF_CIE_DATA_ALIGNMENT
7027 @item DWARF_CIE_DATA_ALIGNMENT
7028 This macro need only be defined if the target might save registers in the
7029 function prologue at an offset to the stack pointer that is not aligned to
7030 @code{UNITS_PER_WORD}. The definition should be the negative minimum
7031 alignment if @code{STACK_GROWS_DOWNWARD} is defined, and the positive
7032 minimum alignment otherwise. @xref{SDB and DWARF}. Only applicable if
7033 the target supports DWARF 2 frame unwind information.
7037 @deftypefn {Target Hook} void TARGET_ASM_EXCEPTION_SECTION ()
7038 If defined, a function that switches to the section in which the main
7039 exception table is to be placed (@pxref{Sections}). The default is a
7040 function that switches to a section named @code{.gcc_except_table} on
7041 machines that support named sections via
7042 @code{TARGET_ASM_NAMED_SECTION}, otherwise if @option{-fpic} or
7043 @option{-fPIC} is in effect, the @code{data_section}, otherwise the
7044 @code{readonly_data_section}.
7047 @deftypefn {Target Hook} void TARGET_ASM_EH_FRAME_SECTION ()
7048 If defined, a function that switches to the section in which the DWARF 2
7049 frame unwind information to be placed (@pxref{Sections}). The default
7050 is a function that outputs a standard GAS section directive, if
7051 @code{EH_FRAME_SECTION_NAME} is defined, or else a data section
7052 directive followed by a synthetic label.
7055 @node Alignment Output
7056 @subsection Assembler Commands for Alignment
7058 @c prevent bad page break with this line
7059 This describes commands for alignment.
7063 @item JUMP_ALIGN (@var{label})
7064 The alignment (log base 2) to put in front of @var{label}, which is
7065 a common destination of jumps and has no fallthru incoming edge.
7067 This macro need not be defined if you don't want any special alignment
7068 to be done at such a time. Most machine descriptions do not currently
7071 Unless it's necessary to inspect the @var{label} parameter, it is better
7072 to set the variable @var{align_jumps} in the target's
7073 @code{OVERRIDE_OPTIONS}. Otherwise, you should try to honor the user's
7074 selection in @var{align_jumps} in a @code{JUMP_ALIGN} implementation.
7076 @findex LABEL_ALIGN_AFTER_BARRIER
7077 @item LABEL_ALIGN_AFTER_BARRIER (@var{label})
7078 The alignment (log base 2) to put in front of @var{label}, which follows
7081 This macro need not be defined if you don't want any special alignment
7082 to be done at such a time. Most machine descriptions do not currently
7085 @findex LABEL_ALIGN_AFTER_BARRIER_MAX_SKIP
7086 @item LABEL_ALIGN_AFTER_BARRIER_MAX_SKIP
7087 The maximum number of bytes to skip when applying
7088 @code{LABEL_ALIGN_AFTER_BARRIER}. This works only if
7089 @code{ASM_OUTPUT_MAX_SKIP_ALIGN} is defined.
7092 @item LOOP_ALIGN (@var{label})
7093 The alignment (log base 2) to put in front of @var{label}, which follows
7094 a @code{NOTE_INSN_LOOP_BEG} note.
7096 This macro need not be defined if you don't want any special alignment
7097 to be done at such a time. Most machine descriptions do not currently
7100 Unless it's necessary to inspect the @var{label} parameter, it is better
7101 to set the variable @code{align_loops} in the target's
7102 @code{OVERRIDE_OPTIONS}. Otherwise, you should try to honor the user's
7103 selection in @code{align_loops} in a @code{LOOP_ALIGN} implementation.
7105 @findex LOOP_ALIGN_MAX_SKIP
7106 @item LOOP_ALIGN_MAX_SKIP
7107 The maximum number of bytes to skip when applying @code{LOOP_ALIGN}.
7108 This works only if @code{ASM_OUTPUT_MAX_SKIP_ALIGN} is defined.
7111 @item LABEL_ALIGN (@var{label})
7112 The alignment (log base 2) to put in front of @var{label}.
7113 If @code{LABEL_ALIGN_AFTER_BARRIER} / @code{LOOP_ALIGN} specify a different alignment,
7114 the maximum of the specified values is used.
7116 Unless it's necessary to inspect the @var{label} parameter, it is better
7117 to set the variable @code{align_labels} in the target's
7118 @code{OVERRIDE_OPTIONS}. Otherwise, you should try to honor the user's
7119 selection in @code{align_labels} in a @code{LABEL_ALIGN} implementation.
7121 @findex LABEL_ALIGN_MAX_SKIP
7122 @item LABEL_ALIGN_MAX_SKIP
7123 The maximum number of bytes to skip when applying @code{LABEL_ALIGN}.
7124 This works only if @code{ASM_OUTPUT_MAX_SKIP_ALIGN} is defined.
7126 @findex ASM_OUTPUT_SKIP
7127 @item ASM_OUTPUT_SKIP (@var{stream}, @var{nbytes})
7128 A C statement to output to the stdio stream @var{stream} an assembler
7129 instruction to advance the location counter by @var{nbytes} bytes.
7130 Those bytes should be zero when loaded. @var{nbytes} will be a C
7131 expression of type @code{int}.
7133 @findex ASM_NO_SKIP_IN_TEXT
7134 @item ASM_NO_SKIP_IN_TEXT
7135 Define this macro if @code{ASM_OUTPUT_SKIP} should not be used in the
7136 text section because it fails to put zeros in the bytes that are skipped.
7137 This is true on many Unix systems, where the pseudo--op to skip bytes
7138 produces no-op instructions rather than zeros when used in the text
7141 @findex ASM_OUTPUT_ALIGN
7142 @item ASM_OUTPUT_ALIGN (@var{stream}, @var{power})
7143 A C statement to output to the stdio stream @var{stream} an assembler
7144 command to advance the location counter to a multiple of 2 to the
7145 @var{power} bytes. @var{power} will be a C expression of type @code{int}.
7147 @findex ASM_OUTPUT_MAX_SKIP_ALIGN
7148 @item ASM_OUTPUT_MAX_SKIP_ALIGN (@var{stream}, @var{power}, @var{max_skip})
7149 A C statement to output to the stdio stream @var{stream} an assembler
7150 command to advance the location counter to a multiple of 2 to the
7151 @var{power} bytes, but only if @var{max_skip} or fewer bytes are needed to
7152 satisfy the alignment request. @var{power} and @var{max_skip} will be
7153 a C expression of type @code{int}.
7157 @node Debugging Info
7158 @section Controlling Debugging Information Format
7160 @c prevent bad page break with this line
7161 This describes how to specify debugging information.
7164 * All Debuggers:: Macros that affect all debugging formats uniformly.
7165 * DBX Options:: Macros enabling specific options in DBX format.
7166 * DBX Hooks:: Hook macros for varying DBX format.
7167 * File Names and DBX:: Macros controlling output of file names in DBX format.
7168 * SDB and DWARF:: Macros for SDB (COFF) and DWARF formats.
7169 * VMS Debug:: Macros for VMS debug format.
7173 @subsection Macros Affecting All Debugging Formats
7175 @c prevent bad page break with this line
7176 These macros affect all debugging formats.
7179 @findex DBX_REGISTER_NUMBER
7180 @item DBX_REGISTER_NUMBER (@var{regno})
7181 A C expression that returns the DBX register number for the compiler
7182 register number @var{regno}. In the default macro provided, the value
7183 of this expression will be @var{regno} itself. But sometimes there are
7184 some registers that the compiler knows about and DBX does not, or vice
7185 versa. In such cases, some register may need to have one number in the
7186 compiler and another for DBX@.
7188 If two registers have consecutive numbers inside GCC, and they can be
7189 used as a pair to hold a multiword value, then they @emph{must} have
7190 consecutive numbers after renumbering with @code{DBX_REGISTER_NUMBER}.
7191 Otherwise, debuggers will be unable to access such a pair, because they
7192 expect register pairs to be consecutive in their own numbering scheme.
7194 If you find yourself defining @code{DBX_REGISTER_NUMBER} in way that
7195 does not preserve register pairs, then what you must do instead is
7196 redefine the actual register numbering scheme.
7198 @findex DEBUGGER_AUTO_OFFSET
7199 @item DEBUGGER_AUTO_OFFSET (@var{x})
7200 A C expression that returns the integer offset value for an automatic
7201 variable having address @var{x} (an RTL expression). The default
7202 computation assumes that @var{x} is based on the frame-pointer and
7203 gives the offset from the frame-pointer. This is required for targets
7204 that produce debugging output for DBX or COFF-style debugging output
7205 for SDB and allow the frame-pointer to be eliminated when the
7206 @option{-g} options is used.
7208 @findex DEBUGGER_ARG_OFFSET
7209 @item DEBUGGER_ARG_OFFSET (@var{offset}, @var{x})
7210 A C expression that returns the integer offset value for an argument
7211 having address @var{x} (an RTL expression). The nominal offset is
7214 @findex PREFERRED_DEBUGGING_TYPE
7215 @item PREFERRED_DEBUGGING_TYPE
7216 A C expression that returns the type of debugging output GCC should
7217 produce when the user specifies just @option{-g}. Define
7218 this if you have arranged for GCC to support more than one format of
7219 debugging output. Currently, the allowable values are @code{DBX_DEBUG},
7220 @code{SDB_DEBUG}, @code{DWARF_DEBUG}, @code{DWARF2_DEBUG},
7221 @code{XCOFF_DEBUG}, @code{VMS_DEBUG}, and @code{VMS_AND_DWARF2_DEBUG}.
7223 When the user specifies @option{-ggdb}, GCC normally also uses the
7224 value of this macro to select the debugging output format, but with two
7225 exceptions. If @code{DWARF2_DEBUGGING_INFO} is defined and
7226 @code{LINKER_DOES_NOT_WORK_WITH_DWARF2} is not defined, GCC uses the
7227 value @code{DWARF2_DEBUG}. Otherwise, if @code{DBX_DEBUGGING_INFO} is
7228 defined, GCC uses @code{DBX_DEBUG}.
7230 The value of this macro only affects the default debugging output; the
7231 user can always get a specific type of output by using @option{-gstabs},
7232 @option{-gcoff}, @option{-gdwarf-1}, @option{-gdwarf-2}, @option{-gxcoff},
7237 @subsection Specific Options for DBX Output
7239 @c prevent bad page break with this line
7240 These are specific options for DBX output.
7243 @findex DBX_DEBUGGING_INFO
7244 @item DBX_DEBUGGING_INFO
7245 Define this macro if GCC should produce debugging output for DBX
7246 in response to the @option{-g} option.
7248 @findex XCOFF_DEBUGGING_INFO
7249 @item XCOFF_DEBUGGING_INFO
7250 Define this macro if GCC should produce XCOFF format debugging output
7251 in response to the @option{-g} option. This is a variant of DBX format.
7253 @findex DEFAULT_GDB_EXTENSIONS
7254 @item DEFAULT_GDB_EXTENSIONS
7255 Define this macro to control whether GCC should by default generate
7256 GDB's extended version of DBX debugging information (assuming DBX-format
7257 debugging information is enabled at all). If you don't define the
7258 macro, the default is 1: always generate the extended information
7259 if there is any occasion to.
7261 @findex DEBUG_SYMS_TEXT
7262 @item DEBUG_SYMS_TEXT
7263 Define this macro if all @code{.stabs} commands should be output while
7264 in the text section.
7266 @findex ASM_STABS_OP
7268 A C string constant, including spacing, naming the assembler pseudo op to
7269 use instead of @code{"\t.stabs\t"} to define an ordinary debugging symbol.
7270 If you don't define this macro, @code{"\t.stabs\t"} is used. This macro
7271 applies only to DBX debugging information format.
7273 @findex ASM_STABD_OP
7275 A C string constant, including spacing, naming the assembler pseudo op to
7276 use instead of @code{"\t.stabd\t"} to define a debugging symbol whose
7277 value is the current location. If you don't define this macro,
7278 @code{"\t.stabd\t"} is used. This macro applies only to DBX debugging
7281 @findex ASM_STABN_OP
7283 A C string constant, including spacing, naming the assembler pseudo op to
7284 use instead of @code{"\t.stabn\t"} to define a debugging symbol with no
7285 name. If you don't define this macro, @code{"\t.stabn\t"} is used. This
7286 macro applies only to DBX debugging information format.
7288 @findex DBX_NO_XREFS
7290 Define this macro if DBX on your system does not support the construct
7291 @samp{xs@var{tagname}}. On some systems, this construct is used to
7292 describe a forward reference to a structure named @var{tagname}.
7293 On other systems, this construct is not supported at all.
7295 @findex DBX_CONTIN_LENGTH
7296 @item DBX_CONTIN_LENGTH
7297 A symbol name in DBX-format debugging information is normally
7298 continued (split into two separate @code{.stabs} directives) when it
7299 exceeds a certain length (by default, 80 characters). On some
7300 operating systems, DBX requires this splitting; on others, splitting
7301 must not be done. You can inhibit splitting by defining this macro
7302 with the value zero. You can override the default splitting-length by
7303 defining this macro as an expression for the length you desire.
7305 @findex DBX_CONTIN_CHAR
7306 @item DBX_CONTIN_CHAR
7307 Normally continuation is indicated by adding a @samp{\} character to
7308 the end of a @code{.stabs} string when a continuation follows. To use
7309 a different character instead, define this macro as a character
7310 constant for the character you want to use. Do not define this macro
7311 if backslash is correct for your system.
7313 @findex DBX_STATIC_STAB_DATA_SECTION
7314 @item DBX_STATIC_STAB_DATA_SECTION
7315 Define this macro if it is necessary to go to the data section before
7316 outputting the @samp{.stabs} pseudo-op for a non-global static
7319 @findex DBX_TYPE_DECL_STABS_CODE
7320 @item DBX_TYPE_DECL_STABS_CODE
7321 The value to use in the ``code'' field of the @code{.stabs} directive
7322 for a typedef. The default is @code{N_LSYM}.
7324 @findex DBX_STATIC_CONST_VAR_CODE
7325 @item DBX_STATIC_CONST_VAR_CODE
7326 The value to use in the ``code'' field of the @code{.stabs} directive
7327 for a static variable located in the text section. DBX format does not
7328 provide any ``right'' way to do this. The default is @code{N_FUN}.
7330 @findex DBX_REGPARM_STABS_CODE
7331 @item DBX_REGPARM_STABS_CODE
7332 The value to use in the ``code'' field of the @code{.stabs} directive
7333 for a parameter passed in registers. DBX format does not provide any
7334 ``right'' way to do this. The default is @code{N_RSYM}.
7336 @findex DBX_REGPARM_STABS_LETTER
7337 @item DBX_REGPARM_STABS_LETTER
7338 The letter to use in DBX symbol data to identify a symbol as a parameter
7339 passed in registers. DBX format does not customarily provide any way to
7340 do this. The default is @code{'P'}.
7342 @findex DBX_MEMPARM_STABS_LETTER
7343 @item DBX_MEMPARM_STABS_LETTER
7344 The letter to use in DBX symbol data to identify a symbol as a stack
7345 parameter. The default is @code{'p'}.
7347 @findex DBX_FUNCTION_FIRST
7348 @item DBX_FUNCTION_FIRST
7349 Define this macro if the DBX information for a function and its
7350 arguments should precede the assembler code for the function. Normally,
7351 in DBX format, the debugging information entirely follows the assembler
7354 @findex DBX_LBRAC_FIRST
7355 @item DBX_LBRAC_FIRST
7356 Define this macro if the @code{N_LBRAC} symbol for a block should
7357 precede the debugging information for variables and functions defined in
7358 that block. Normally, in DBX format, the @code{N_LBRAC} symbol comes
7361 @findex DBX_BLOCKS_FUNCTION_RELATIVE
7362 @item DBX_BLOCKS_FUNCTION_RELATIVE
7363 Define this macro if the value of a symbol describing the scope of a
7364 block (@code{N_LBRAC} or @code{N_RBRAC}) should be relative to the start
7365 of the enclosing function. Normally, GCC uses an absolute address.
7367 @findex DBX_USE_BINCL
7369 Define this macro if GCC should generate @code{N_BINCL} and
7370 @code{N_EINCL} stabs for included header files, as on Sun systems. This
7371 macro also directs GCC to output a type number as a pair of a file
7372 number and a type number within the file. Normally, GCC does not
7373 generate @code{N_BINCL} or @code{N_EINCL} stabs, and it outputs a single
7374 number for a type number.
7378 @subsection Open-Ended Hooks for DBX Format
7380 @c prevent bad page break with this line
7381 These are hooks for DBX format.
7384 @findex DBX_OUTPUT_LBRAC
7385 @item DBX_OUTPUT_LBRAC (@var{stream}, @var{name})
7386 Define this macro to say how to output to @var{stream} the debugging
7387 information for the start of a scope level for variable names. The
7388 argument @var{name} is the name of an assembler symbol (for use with
7389 @code{assemble_name}) whose value is the address where the scope begins.
7391 @findex DBX_OUTPUT_RBRAC
7392 @item DBX_OUTPUT_RBRAC (@var{stream}, @var{name})
7393 Like @code{DBX_OUTPUT_LBRAC}, but for the end of a scope level.
7395 @findex DBX_OUTPUT_ENUM
7396 @item DBX_OUTPUT_ENUM (@var{stream}, @var{type})
7397 Define this macro if the target machine requires special handling to
7398 output an enumeration type. The definition should be a C statement
7399 (sans semicolon) to output the appropriate information to @var{stream}
7400 for the type @var{type}.
7402 @findex DBX_OUTPUT_FUNCTION_END
7403 @item DBX_OUTPUT_FUNCTION_END (@var{stream}, @var{function})
7404 Define this macro if the target machine requires special output at the
7405 end of the debugging information for a function. The definition should
7406 be a C statement (sans semicolon) to output the appropriate information
7407 to @var{stream}. @var{function} is the @code{FUNCTION_DECL} node for
7410 @findex DBX_OUTPUT_STANDARD_TYPES
7411 @item DBX_OUTPUT_STANDARD_TYPES (@var{syms})
7412 Define this macro if you need to control the order of output of the
7413 standard data types at the beginning of compilation. The argument
7414 @var{syms} is a @code{tree} which is a chain of all the predefined
7415 global symbols, including names of data types.
7417 Normally, DBX output starts with definitions of the types for integers
7418 and characters, followed by all the other predefined types of the
7419 particular language in no particular order.
7421 On some machines, it is necessary to output different particular types
7422 first. To do this, define @code{DBX_OUTPUT_STANDARD_TYPES} to output
7423 those symbols in the necessary order. Any predefined types that you
7424 don't explicitly output will be output afterward in no particular order.
7426 Be careful not to define this macro so that it works only for C@. There
7427 are no global variables to access most of the built-in types, because
7428 another language may have another set of types. The way to output a
7429 particular type is to look through @var{syms} to see if you can find it.
7435 for (decl = syms; decl; decl = TREE_CHAIN (decl))
7436 if (!strcmp (IDENTIFIER_POINTER (DECL_NAME (decl)),
7438 dbxout_symbol (decl);
7444 This does nothing if the expected type does not exist.
7446 See the function @code{init_decl_processing} in @file{c-decl.c} to find
7447 the names to use for all the built-in C types.
7449 Here is another way of finding a particular type:
7451 @c this is still overfull. --mew 10feb93
7455 for (decl = syms; decl; decl = TREE_CHAIN (decl))
7456 if (TREE_CODE (decl) == TYPE_DECL
7457 && (TREE_CODE (TREE_TYPE (decl))
7459 && TYPE_PRECISION (TREE_TYPE (decl)) == 16
7460 && TYPE_UNSIGNED (TREE_TYPE (decl)))
7462 /* @r{This must be @code{unsigned short}.} */
7463 dbxout_symbol (decl);
7469 @findex NO_DBX_FUNCTION_END
7470 @item NO_DBX_FUNCTION_END
7471 Some stabs encapsulation formats (in particular ECOFF), cannot handle the
7472 @code{.stabs "",N_FUN,,0,0,Lscope-function-1} gdb dbx extension construct.
7473 On those machines, define this macro to turn this feature off without
7474 disturbing the rest of the gdb extensions.
7478 @node File Names and DBX
7479 @subsection File Names in DBX Format
7481 @c prevent bad page break with this line
7482 This describes file names in DBX format.
7485 @findex DBX_WORKING_DIRECTORY
7486 @item DBX_WORKING_DIRECTORY
7487 Define this if DBX wants to have the current directory recorded in each
7490 Note that the working directory is always recorded if GDB extensions are
7493 @findex DBX_OUTPUT_MAIN_SOURCE_FILENAME
7494 @item DBX_OUTPUT_MAIN_SOURCE_FILENAME (@var{stream}, @var{name})
7495 A C statement to output DBX debugging information to the stdio stream
7496 @var{stream} which indicates that file @var{name} is the main source
7497 file---the file specified as the input file for compilation.
7498 This macro is called only once, at the beginning of compilation.
7500 This macro need not be defined if the standard form of output
7501 for DBX debugging information is appropriate.
7503 @findex DBX_OUTPUT_MAIN_SOURCE_DIRECTORY
7504 @item DBX_OUTPUT_MAIN_SOURCE_DIRECTORY (@var{stream}, @var{name})
7505 A C statement to output DBX debugging information to the stdio stream
7506 @var{stream} which indicates that the current directory during
7507 compilation is named @var{name}.
7509 This macro need not be defined if the standard form of output
7510 for DBX debugging information is appropriate.
7512 @findex DBX_OUTPUT_MAIN_SOURCE_FILE_END
7513 @item DBX_OUTPUT_MAIN_SOURCE_FILE_END (@var{stream}, @var{name})
7514 A C statement to output DBX debugging information at the end of
7515 compilation of the main source file @var{name}.
7517 If you don't define this macro, nothing special is output at the end
7518 of compilation, which is correct for most machines.
7520 @findex DBX_OUTPUT_SOURCE_FILENAME
7521 @item DBX_OUTPUT_SOURCE_FILENAME (@var{stream}, @var{name})
7522 A C statement to output DBX debugging information to the stdio stream
7523 @var{stream} which indicates that file @var{name} is the current source
7524 file. This output is generated each time input shifts to a different
7525 source file as a result of @samp{#include}, the end of an included file,
7526 or a @samp{#line} command.
7528 This macro need not be defined if the standard form of output
7529 for DBX debugging information is appropriate.
7534 @subsection Macros for SDB and DWARF Output
7536 @c prevent bad page break with this line
7537 Here are macros for SDB and DWARF output.
7540 @findex SDB_DEBUGGING_INFO
7541 @item SDB_DEBUGGING_INFO
7542 Define this macro if GCC should produce COFF-style debugging output
7543 for SDB in response to the @option{-g} option.
7545 @findex DWARF_DEBUGGING_INFO
7546 @item DWARF_DEBUGGING_INFO
7547 Define this macro if GCC should produce dwarf format debugging output
7548 in response to the @option{-g} option.
7550 @findex DWARF2_DEBUGGING_INFO
7551 @item DWARF2_DEBUGGING_INFO
7552 Define this macro if GCC should produce dwarf version 2 format
7553 debugging output in response to the @option{-g} option.
7555 To support optional call frame debugging information, you must also
7556 define @code{INCOMING_RETURN_ADDR_RTX} and either set
7557 @code{RTX_FRAME_RELATED_P} on the prologue insns if you use RTL for the
7558 prologue, or call @code{dwarf2out_def_cfa} and @code{dwarf2out_reg_save}
7559 as appropriate from @code{TARGET_ASM_FUNCTION_PROLOGUE} if you don't.
7561 @findex DWARF2_FRAME_INFO
7562 @item DWARF2_FRAME_INFO
7563 Define this macro to a nonzero value if GCC should always output
7564 Dwarf 2 frame information. If @code{DWARF2_UNWIND_INFO}
7565 (@pxref{Exception Region Output} is nonzero, GCC will output this
7566 information not matter how you define @code{DWARF2_FRAME_INFO}.
7568 @findex LINKER_DOES_NOT_WORK_WITH_DWARF2
7569 @item LINKER_DOES_NOT_WORK_WITH_DWARF2
7570 Define this macro if the linker does not work with Dwarf version 2.
7571 Normally, if the user specifies only @option{-ggdb} GCC will use Dwarf
7572 version 2 if available; this macro disables this. See the description
7573 of the @code{PREFERRED_DEBUGGING_TYPE} macro for more details.
7575 @findex DWARF2_GENERATE_TEXT_SECTION_LABEL
7576 @item DWARF2_GENERATE_TEXT_SECTION_LABEL
7577 By default, the Dwarf 2 debugging information generator will generate a
7578 label to mark the beginning of the text section. If it is better simply
7579 to use the name of the text section itself, rather than an explicit label,
7580 to indicate the beginning of the text section, define this macro to zero.
7582 @findex DWARF2_ASM_LINE_DEBUG_INFO
7583 @item DWARF2_ASM_LINE_DEBUG_INFO
7584 Define this macro to be a nonzero value if the assembler can generate Dwarf 2
7585 line debug info sections. This will result in much more compact line number
7586 tables, and hence is desirable if it works.
7588 @findex PUT_SDB_@dots{}
7589 @item PUT_SDB_@dots{}
7590 Define these macros to override the assembler syntax for the special
7591 SDB assembler directives. See @file{sdbout.c} for a list of these
7592 macros and their arguments. If the standard syntax is used, you need
7593 not define them yourself.
7597 Some assemblers do not support a semicolon as a delimiter, even between
7598 SDB assembler directives. In that case, define this macro to be the
7599 delimiter to use (usually @samp{\n}). It is not necessary to define
7600 a new set of @code{PUT_SDB_@var{op}} macros if this is the only change
7603 @findex SDB_GENERATE_FAKE
7604 @item SDB_GENERATE_FAKE
7605 Define this macro to override the usual method of constructing a dummy
7606 name for anonymous structure and union types. See @file{sdbout.c} for
7609 @findex SDB_ALLOW_UNKNOWN_REFERENCES
7610 @item SDB_ALLOW_UNKNOWN_REFERENCES
7611 Define this macro to allow references to unknown structure,
7612 union, or enumeration tags to be emitted. Standard COFF does not
7613 allow handling of unknown references, MIPS ECOFF has support for
7616 @findex SDB_ALLOW_FORWARD_REFERENCES
7617 @item SDB_ALLOW_FORWARD_REFERENCES
7618 Define this macro to allow references to structure, union, or
7619 enumeration tags that have not yet been seen to be handled. Some
7620 assemblers choke if forward tags are used, while some require it.
7625 @subsection Macros for VMS Debug Format
7627 @c prevent bad page break with this line
7628 Here are macros for VMS debug format.
7631 @findex VMS_DEBUGGING_INFO
7632 @item VMS_DEBUGGING_INFO
7633 Define this macro if GCC should produce debugging output for VMS
7634 in response to the @option{-g} option. The default behavior for VMS
7635 is to generate minimal debug info for a traceback in the absence of
7636 @option{-g} unless explicitly overridden with @option{-g0}. This
7637 behavior is controlled by @code{OPTIMIZATION_OPTIONS} and
7638 @code{OVERRIDE_OPTIONS}.
7641 @node Cross-compilation
7642 @section Cross Compilation and Floating Point
7643 @cindex cross compilation and floating point
7644 @cindex floating point and cross compilation
7646 While all modern machines use 2's complement representation for integers,
7647 there are a variety of representations for floating point numbers. This
7648 means that in a cross-compiler the representation of floating point numbers
7649 in the compiled program may be different from that used in the machine
7650 doing the compilation.
7653 Because different representation systems may offer different amounts of
7654 range and precision, the cross compiler cannot safely use the host
7655 machine's floating point arithmetic. Therefore, floating point constants
7656 must be represented in the target machine's format. This means that the
7657 cross compiler cannot use @code{atof} to parse a floating point constant;
7658 it must have its own special routine to use instead. Also, constant
7659 folding must emulate the target machine's arithmetic (or must not be done
7662 The macros in the following table should be defined only if you are cross
7663 compiling between different floating point formats.
7665 Otherwise, don't define them. Then default definitions will be set up which
7666 use @code{double} as the data type, @code{==} to test for equality, etc.
7668 You don't need to worry about how many times you use an operand of any
7669 of these macros. The compiler never uses operands which have side effects.
7672 @findex REAL_VALUE_TYPE
7673 @item REAL_VALUE_TYPE
7674 A macro for the C data type to be used to hold a floating point value
7675 in the target machine's format. Typically this would be a
7676 @code{struct} containing an array of @code{int}.
7678 @findex REAL_VALUES_EQUAL
7679 @item REAL_VALUES_EQUAL (@var{x}, @var{y})
7680 A macro for a C expression which compares for equality the two values,
7681 @var{x} and @var{y}, both of type @code{REAL_VALUE_TYPE}.
7683 @findex REAL_VALUES_LESS
7684 @item REAL_VALUES_LESS (@var{x}, @var{y})
7685 A macro for a C expression which tests whether @var{x} is less than
7686 @var{y}, both values being of type @code{REAL_VALUE_TYPE} and
7687 interpreted as floating point numbers in the target machine's
7690 @findex REAL_VALUE_LDEXP
7692 @item REAL_VALUE_LDEXP (@var{x}, @var{scale})
7693 A macro for a C expression which performs the standard library
7694 function @code{ldexp}, but using the target machine's floating point
7695 representation. Both @var{x} and the value of the expression have
7696 type @code{REAL_VALUE_TYPE}. The second argument, @var{scale}, is an
7699 @findex REAL_VALUE_FIX
7700 @item REAL_VALUE_FIX (@var{x})
7701 A macro whose definition is a C expression to convert the target-machine
7702 floating point value @var{x} to a signed integer. @var{x} has type
7703 @code{REAL_VALUE_TYPE}.
7705 @findex REAL_VALUE_UNSIGNED_FIX
7706 @item REAL_VALUE_UNSIGNED_FIX (@var{x})
7707 A macro whose definition is a C expression to convert the target-machine
7708 floating point value @var{x} to an unsigned integer. @var{x} has type
7709 @code{REAL_VALUE_TYPE}.
7711 @findex REAL_VALUE_RNDZINT
7712 @item REAL_VALUE_RNDZINT (@var{x})
7713 A macro whose definition is a C expression to round the target-machine
7714 floating point value @var{x} towards zero to an integer value (but still
7715 as a floating point number). @var{x} has type @code{REAL_VALUE_TYPE},
7716 and so does the value.
7718 @findex REAL_VALUE_UNSIGNED_RNDZINT
7719 @item REAL_VALUE_UNSIGNED_RNDZINT (@var{x})
7720 A macro whose definition is a C expression to round the target-machine
7721 floating point value @var{x} towards zero to an unsigned integer value
7722 (but still represented as a floating point number). @var{x} has type
7723 @code{REAL_VALUE_TYPE}, and so does the value.
7725 @findex REAL_VALUE_ATOF
7726 @item REAL_VALUE_ATOF (@var{string}, @var{mode})
7727 A macro for a C expression which converts @var{string}, an expression of
7728 type @code{char *}, into a floating point number in the target machine's
7729 representation for mode @var{mode}. The value has type
7730 @code{REAL_VALUE_TYPE}.
7732 @findex REAL_INFINITY
7734 Define this macro if infinity is a possible floating point value, and
7735 therefore division by 0 is legitimate.
7737 @findex REAL_VALUE_ISINF
7739 @item REAL_VALUE_ISINF (@var{x})
7740 A macro for a C expression which determines whether @var{x}, a floating
7741 point value, is infinity. The value has type @code{int}.
7742 By default, this is defined to call @code{isinf}.
7744 @findex REAL_VALUE_ISNAN
7746 @item REAL_VALUE_ISNAN (@var{x})
7747 A macro for a C expression which determines whether @var{x}, a floating
7748 point value, is a ``nan'' (not-a-number). The value has type
7749 @code{int}. By default, this is defined to call @code{isnan}.
7752 @cindex constant folding and floating point
7753 Define the following additional macros if you want to make floating
7754 point constant folding work while cross compiling. If you don't
7755 define them, cross compilation is still possible, but constant folding
7756 will not happen for floating point values.
7759 @findex REAL_ARITHMETIC
7760 @item REAL_ARITHMETIC (@var{output}, @var{code}, @var{x}, @var{y})
7761 A macro for a C statement which calculates an arithmetic operation of
7762 the two floating point values @var{x} and @var{y}, both of type
7763 @code{REAL_VALUE_TYPE} in the target machine's representation, to
7764 produce a result of the same type and representation which is stored
7765 in @var{output} (which will be a variable).
7767 The operation to be performed is specified by @var{code}, a tree code
7768 which will always be one of the following: @code{PLUS_EXPR},
7769 @code{MINUS_EXPR}, @code{MULT_EXPR}, @code{RDIV_EXPR},
7770 @code{MAX_EXPR}, @code{MIN_EXPR}.
7772 @cindex overflow while constant folding
7773 The expansion of this macro is responsible for checking for overflow.
7774 If overflow happens, the macro expansion should execute the statement
7775 @code{return 0;}, which indicates the inability to perform the
7776 arithmetic operation requested.
7778 @findex REAL_VALUE_NEGATE
7779 @item REAL_VALUE_NEGATE (@var{x})
7780 A macro for a C expression which returns the negative of the floating
7781 point value @var{x}. Both @var{x} and the value of the expression
7782 have type @code{REAL_VALUE_TYPE} and are in the target machine's
7783 floating point representation.
7785 There is no way for this macro to report overflow, since overflow
7786 can't happen in the negation operation.
7788 @findex REAL_VALUE_TRUNCATE
7789 @item REAL_VALUE_TRUNCATE (@var{mode}, @var{x})
7790 A macro for a C expression which converts the floating point value
7791 @var{x} to mode @var{mode}.
7793 Both @var{x} and the value of the expression are in the target machine's
7794 floating point representation and have type @code{REAL_VALUE_TYPE}.
7795 However, the value should have an appropriate bit pattern to be output
7796 properly as a floating constant whose precision accords with mode
7799 There is no way for this macro to report overflow.
7801 @findex REAL_VALUE_TO_INT
7802 @item REAL_VALUE_TO_INT (@var{low}, @var{high}, @var{x})
7803 A macro for a C expression which converts a floating point value
7804 @var{x} into a double-precision integer which is then stored into
7805 @var{low} and @var{high}, two variables of type @var{int}.
7807 @item REAL_VALUE_FROM_INT (@var{x}, @var{low}, @var{high}, @var{mode})
7808 @findex REAL_VALUE_FROM_INT
7809 A macro for a C expression which converts a double-precision integer
7810 found in @var{low} and @var{high}, two variables of type @var{int},
7811 into a floating point value which is then stored into @var{x}.
7812 The value is in the target machine's representation for mode @var{mode}
7813 and has the type @code{REAL_VALUE_TYPE}.
7816 @node Mode Switching
7817 @section Mode Switching Instructions
7818 @cindex mode switching
7819 The following macros control mode switching optimizations:
7822 @findex OPTIMIZE_MODE_SWITCHING
7823 @item OPTIMIZE_MODE_SWITCHING (@var{entity})
7824 Define this macro if the port needs extra instructions inserted for mode
7825 switching in an optimizing compilation.
7827 For an example, the SH4 can perform both single and double precision
7828 floating point operations, but to perform a single precision operation,
7829 the FPSCR PR bit has to be cleared, while for a double precision
7830 operation, this bit has to be set. Changing the PR bit requires a general
7831 purpose register as a scratch register, hence these FPSCR sets have to
7832 be inserted before reload, i.e.@: you can't put this into instruction emitting
7833 or @code{MACHINE_DEPENDENT_REORG}.
7835 You can have multiple entities that are mode-switched, and select at run time
7836 which entities actually need it. @code{OPTIMIZE_MODE_SWITCHING} should
7837 return nonzero for any @var{entity} that needs mode-switching.
7838 If you define this macro, you also have to define
7839 @code{NUM_MODES_FOR_MODE_SWITCHING}, @code{MODE_NEEDED},
7840 @code{MODE_PRIORITY_TO_MODE} and @code{EMIT_MODE_SET}.
7841 @code{NORMAL_MODE} is optional.
7843 @findex NUM_MODES_FOR_MODE_SWITCHING
7844 @item NUM_MODES_FOR_MODE_SWITCHING
7845 If you define @code{OPTIMIZE_MODE_SWITCHING}, you have to define this as
7846 initializer for an array of integers. Each initializer element
7847 N refers to an entity that needs mode switching, and specifies the number
7848 of different modes that might need to be set for this entity.
7849 The position of the initializer in the initializer - starting counting at
7850 zero - determines the integer that is used to refer to the mode-switched
7852 In macros that take mode arguments / yield a mode result, modes are
7853 represented as numbers 0 @dots{} N @minus{} 1. N is used to specify that no mode
7854 switch is needed / supplied.
7857 @item MODE_NEEDED (@var{entity}, @var{insn})
7858 @var{entity} is an integer specifying a mode-switched entity. If
7859 @code{OPTIMIZE_MODE_SWITCHING} is defined, you must define this macro to
7860 return an integer value not larger than the corresponding element in
7861 @code{NUM_MODES_FOR_MODE_SWITCHING}, to denote the mode that @var{entity} must
7862 be switched into prior to the execution of @var{insn}.
7865 @item NORMAL_MODE (@var{entity})
7866 If this macro is defined, it is evaluated for every @var{entity} that needs
7867 mode switching. It should evaluate to an integer, which is a mode that
7868 @var{entity} is assumed to be switched to at function entry and exit.
7870 @findex MODE_PRIORITY_TO_MODE
7871 @item MODE_PRIORITY_TO_MODE (@var{entity}, @var{n})
7872 This macro specifies the order in which modes for @var{entity} are processed.
7873 0 is the highest priority, @code{NUM_MODES_FOR_MODE_SWITCHING[@var{entity}] - 1} the
7874 lowest. The value of the macro should be an integer designating a mode
7875 for @var{entity}. For any fixed @var{entity}, @code{mode_priority_to_mode}
7876 (@var{entity}, @var{n}) shall be a bijection in 0 @dots{}
7877 @code{num_modes_for_mode_switching[@var{entity}] - 1}.
7879 @findex EMIT_MODE_SET
7880 @item EMIT_MODE_SET (@var{entity}, @var{mode}, @var{hard_regs_live})
7881 Generate one or more insns to set @var{entity} to @var{mode}.
7882 @var{hard_reg_live} is the set of hard registers live at the point where
7883 the insn(s) are to be inserted.
7886 @node Target Attributes
7887 @section Defining target-specific uses of @code{__attribute__}
7888 @cindex target attributes
7889 @cindex machine attributes
7890 @cindex attributes, target-specific
7892 Target-specific attributes may be defined for functions, data and types.
7893 These are described using the following target hooks; they also need to
7894 be documented in @file{extend.texi}.
7896 @deftypevr {Target Hook} {const struct attribute_spec *} TARGET_ATTRIBUTE_TABLE
7897 If defined, this target hook points to an array of @samp{struct
7898 attribute_spec} (defined in @file{tree.h}) specifying the machine
7899 specific attributes for this target and some of the restrictions on the
7900 entities to which these attributes are applied and the arguments they
7904 @deftypefn {Target Hook} int TARGET_COMP_TYPE_ATTRIBUTES (tree @var{type1}, tree @var{type2})
7905 If defined, this target hook is a function which returns zero if the attributes on
7906 @var{type1} and @var{type2} are incompatible, one if they are compatible,
7907 and two if they are nearly compatible (which causes a warning to be
7908 generated). If this is not defined, machine-specific attributes are
7909 supposed always to be compatible.
7912 @deftypefn {Target Hook} void TARGET_SET_DEFAULT_TYPE_ATTRIBUTES (tree @var{type})
7913 If defined, this target hook is a function which assigns default attributes to
7914 newly defined @var{type}.
7917 @deftypefn {Target Hook} tree TARGET_MERGE_TYPE_ATTRIBUTES (tree @var{type1}, tree @var{type2})
7918 Define this target hook if the merging of type attributes needs special
7919 handling. If defined, the result is a list of the combined
7920 @code{TYPE_ATTRIBUTES} of @var{type1} and @var{type2}. It is assumed
7921 that @code{comptypes} has already been called and returned 1. This
7922 function may call @code{merge_attributes} to handle machine-independent
7926 @deftypefn {Target Hook} tree TARGET_MERGE_DECL_ATTRIBUTES (tree @var{olddecl}, tree @var{newdecl})
7927 Define this target hook if the merging of decl attributes needs special
7928 handling. If defined, the result is a list of the combined
7929 @code{DECL_ATTRIBUTES} of @var{olddecl} and @var{newdecl}.
7930 @var{newdecl} is a duplicate declaration of @var{olddecl}. Examples of
7931 when this is needed are when one attribute overrides another, or when an
7932 attribute is nullified by a subsequent definition. This function may
7933 call @code{merge_attributes} to handle machine-independent merging.
7935 @findex TARGET_DLLIMPORT_DECL_ATTRIBUTES
7936 If the only target-specific handling you require is @samp{dllimport} for
7937 Windows targets, you should define the macro
7938 @code{TARGET_DLLIMPORT_DECL_ATTRIBUTES}. This links in a function
7939 called @code{merge_dllimport_decl_attributes} which can then be defined
7940 as the expansion of @code{TARGET_MERGE_DECL_ATTRIBUTES}. This is done
7941 in @file{i386/cygwin.h} and @file{i386/i386.c}, for example.
7944 @deftypefn {Target Hook} void TARGET_INSERT_ATTRIBUTES (tree @var{node}, tree *@var{attr_ptr})
7945 Define this target hook if you want to be able to add attributes to a decl
7946 when it is being created. This is normally useful for back ends which
7947 wish to implement a pragma by using the attributes which correspond to
7948 the pragma's effect. The @var{node} argument is the decl which is being
7949 created. The @var{attr_ptr} argument is a pointer to the attribute list
7950 for this decl. The list itself should not be modified, since it may be
7951 shared with other decls, but attributes may be chained on the head of
7952 the list and @code{*@var{attr_ptr}} modified to point to the new
7953 attributes, or a copy of the list may be made if further changes are
7957 @deftypefn {Target Hook} bool TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P (tree @var{fndecl})
7959 This target hook returns @code{true} if it is ok to inline @var{fndecl}
7960 into the current function, despite its having target-specific
7961 attributes, @code{false} otherwise. By default, if a function has a
7962 target specific attribute attached to it, it will not be inlined.
7966 @section Miscellaneous Parameters
7967 @cindex parameters, miscellaneous
7969 @c prevent bad page break with this line
7970 Here are several miscellaneous parameters.
7973 @item PREDICATE_CODES
7974 @findex PREDICATE_CODES
7975 Define this if you have defined special-purpose predicates in the file
7976 @file{@var{machine}.c}. This macro is called within an initializer of an
7977 array of structures. The first field in the structure is the name of a
7978 predicate and the second field is an array of rtl codes. For each
7979 predicate, list all rtl codes that can be in expressions matched by the
7980 predicate. The list should have a trailing comma. Here is an example
7981 of two entries in the list for a typical RISC machine:
7984 #define PREDICATE_CODES \
7985 @{"gen_reg_rtx_operand", @{SUBREG, REG@}@}, \
7986 @{"reg_or_short_cint_operand", @{SUBREG, REG, CONST_INT@}@},
7989 Defining this macro does not affect the generated code (however,
7990 incorrect definitions that omit an rtl code that may be matched by the
7991 predicate can cause the compiler to malfunction). Instead, it allows
7992 the table built by @file{genrecog} to be more compact and efficient,
7993 thus speeding up the compiler. The most important predicates to include
7994 in the list specified by this macro are those used in the most insn
7997 For each predicate function named in @code{PREDICATE_CODES}, a
7998 declaration will be generated in @file{insn-codes.h}.
8000 @item SPECIAL_MODE_PREDICATES
8001 @findex SPECIAL_MODE_PREDICATES
8002 Define this if you have special predicates that know special things
8003 about modes. Genrecog will warn about certain forms of
8004 @code{match_operand} without a mode; if the operand predicate is
8005 listed in @code{SPECIAL_MODE_PREDICATES}, the warning will be
8008 Here is an example from the IA-32 port (@code{ext_register_operand}
8009 specially checks for @code{HImode} or @code{SImode} in preparation
8010 for a byte extraction from @code{%ah} etc.).
8013 #define SPECIAL_MODE_PREDICATES \
8014 "ext_register_operand",
8017 @findex CASE_VECTOR_MODE
8018 @item CASE_VECTOR_MODE
8019 An alias for a machine mode name. This is the machine mode that
8020 elements of a jump-table should have.
8022 @findex CASE_VECTOR_SHORTEN_MODE
8023 @item CASE_VECTOR_SHORTEN_MODE (@var{min_offset}, @var{max_offset}, @var{body})
8024 Optional: return the preferred mode for an @code{addr_diff_vec}
8025 when the minimum and maximum offset are known. If you define this,
8026 it enables extra code in branch shortening to deal with @code{addr_diff_vec}.
8027 To make this work, you also have to define INSN_ALIGN and
8028 make the alignment for @code{addr_diff_vec} explicit.
8029 The @var{body} argument is provided so that the offset_unsigned and scale
8030 flags can be updated.
8032 @findex CASE_VECTOR_PC_RELATIVE
8033 @item CASE_VECTOR_PC_RELATIVE
8034 Define this macro to be a C expression to indicate when jump-tables
8035 should contain relative addresses. If jump-tables never contain
8036 relative addresses, then you need not define this macro.
8038 @findex CASE_DROPS_THROUGH
8039 @item CASE_DROPS_THROUGH
8040 Define this if control falls through a @code{case} insn when the index
8041 value is out of range. This means the specified default-label is
8042 actually ignored by the @code{case} insn proper.
8044 @findex CASE_VALUES_THRESHOLD
8045 @item CASE_VALUES_THRESHOLD
8046 Define this to be the smallest number of different values for which it
8047 is best to use a jump-table instead of a tree of conditional branches.
8048 The default is four for machines with a @code{casesi} instruction and
8049 five otherwise. This is best for most machines.
8051 @findex WORD_REGISTER_OPERATIONS
8052 @item WORD_REGISTER_OPERATIONS
8053 Define this macro if operations between registers with integral mode
8054 smaller than a word are always performed on the entire register.
8055 Most RISC machines have this property and most CISC machines do not.
8057 @findex LOAD_EXTEND_OP
8058 @item LOAD_EXTEND_OP (@var{mode})
8059 Define this macro to be a C expression indicating when insns that read
8060 memory in @var{mode}, an integral mode narrower than a word, set the
8061 bits outside of @var{mode} to be either the sign-extension or the
8062 zero-extension of the data read. Return @code{SIGN_EXTEND} for values
8063 of @var{mode} for which the
8064 insn sign-extends, @code{ZERO_EXTEND} for which it zero-extends, and
8065 @code{NIL} for other modes.
8067 This macro is not called with @var{mode} non-integral or with a width
8068 greater than or equal to @code{BITS_PER_WORD}, so you may return any
8069 value in this case. Do not define this macro if it would always return
8070 @code{NIL}. On machines where this macro is defined, you will normally
8071 define it as the constant @code{SIGN_EXTEND} or @code{ZERO_EXTEND}.
8073 @findex SHORT_IMMEDIATES_SIGN_EXTEND
8074 @item SHORT_IMMEDIATES_SIGN_EXTEND
8075 Define this macro if loading short immediate values into registers sign
8078 @findex FIXUNS_TRUNC_LIKE_FIX_TRUNC
8079 @item FIXUNS_TRUNC_LIKE_FIX_TRUNC
8080 Define this macro if the same instructions that convert a floating
8081 point number to a signed fixed point number also convert validly to an
8086 The maximum number of bytes that a single instruction can move quickly
8087 between memory and registers or between two memory locations.
8089 @findex MAX_MOVE_MAX
8091 The maximum number of bytes that a single instruction can move quickly
8092 between memory and registers or between two memory locations. If this
8093 is undefined, the default is @code{MOVE_MAX}. Otherwise, it is the
8094 constant value that is the largest value that @code{MOVE_MAX} can have
8097 @findex SHIFT_COUNT_TRUNCATED
8098 @item SHIFT_COUNT_TRUNCATED
8099 A C expression that is nonzero if on this machine the number of bits
8100 actually used for the count of a shift operation is equal to the number
8101 of bits needed to represent the size of the object being shifted. When
8102 this macro is nonzero, the compiler will assume that it is safe to omit
8103 a sign-extend, zero-extend, and certain bitwise `and' instructions that
8104 truncates the count of a shift operation. On machines that have
8105 instructions that act on bit-fields at variable positions, which may
8106 include `bit test' instructions, a nonzero @code{SHIFT_COUNT_TRUNCATED}
8107 also enables deletion of truncations of the values that serve as
8108 arguments to bit-field instructions.
8110 If both types of instructions truncate the count (for shifts) and
8111 position (for bit-field operations), or if no variable-position bit-field
8112 instructions exist, you should define this macro.
8114 However, on some machines, such as the 80386 and the 680x0, truncation
8115 only applies to shift operations and not the (real or pretended)
8116 bit-field operations. Define @code{SHIFT_COUNT_TRUNCATED} to be zero on
8117 such machines. Instead, add patterns to the @file{md} file that include
8118 the implied truncation of the shift instructions.
8120 You need not define this macro if it would always have the value of zero.
8122 @findex TRULY_NOOP_TRUNCATION
8123 @item TRULY_NOOP_TRUNCATION (@var{outprec}, @var{inprec})
8124 A C expression which is nonzero if on this machine it is safe to
8125 ``convert'' an integer of @var{inprec} bits to one of @var{outprec}
8126 bits (where @var{outprec} is smaller than @var{inprec}) by merely
8127 operating on it as if it had only @var{outprec} bits.
8129 On many machines, this expression can be 1.
8131 @c rearranged this, removed the phrase "it is reported that". this was
8132 @c to fix an overfull hbox. --mew 10feb93
8133 When @code{TRULY_NOOP_TRUNCATION} returns 1 for a pair of sizes for
8134 modes for which @code{MODES_TIEABLE_P} is 0, suboptimal code can result.
8135 If this is the case, making @code{TRULY_NOOP_TRUNCATION} return 0 in
8136 such cases may improve things.
8138 @findex STORE_FLAG_VALUE
8139 @item STORE_FLAG_VALUE
8140 A C expression describing the value returned by a comparison operator
8141 with an integral mode and stored by a store-flag instruction
8142 (@samp{s@var{cond}}) when the condition is true. This description must
8143 apply to @emph{all} the @samp{s@var{cond}} patterns and all the
8144 comparison operators whose results have a @code{MODE_INT} mode.
8146 A value of 1 or @minus{}1 means that the instruction implementing the
8147 comparison operator returns exactly 1 or @minus{}1 when the comparison is true
8148 and 0 when the comparison is false. Otherwise, the value indicates
8149 which bits of the result are guaranteed to be 1 when the comparison is
8150 true. This value is interpreted in the mode of the comparison
8151 operation, which is given by the mode of the first operand in the
8152 @samp{s@var{cond}} pattern. Either the low bit or the sign bit of
8153 @code{STORE_FLAG_VALUE} be on. Presently, only those bits are used by
8156 If @code{STORE_FLAG_VALUE} is neither 1 or @minus{}1, the compiler will
8157 generate code that depends only on the specified bits. It can also
8158 replace comparison operators with equivalent operations if they cause
8159 the required bits to be set, even if the remaining bits are undefined.
8160 For example, on a machine whose comparison operators return an
8161 @code{SImode} value and where @code{STORE_FLAG_VALUE} is defined as
8162 @samp{0x80000000}, saying that just the sign bit is relevant, the
8166 (ne:SI (and:SI @var{x} (const_int @var{power-of-2})) (const_int 0))
8173 (ashift:SI @var{x} (const_int @var{n}))
8177 where @var{n} is the appropriate shift count to move the bit being
8178 tested into the sign bit.
8180 There is no way to describe a machine that always sets the low-order bit
8181 for a true value, but does not guarantee the value of any other bits,
8182 but we do not know of any machine that has such an instruction. If you
8183 are trying to port GCC to such a machine, include an instruction to
8184 perform a logical-and of the result with 1 in the pattern for the
8185 comparison operators and let us know at @email{gcc@@gcc.gnu.org}.
8187 Often, a machine will have multiple instructions that obtain a value
8188 from a comparison (or the condition codes). Here are rules to guide the
8189 choice of value for @code{STORE_FLAG_VALUE}, and hence the instructions
8194 Use the shortest sequence that yields a valid definition for
8195 @code{STORE_FLAG_VALUE}. It is more efficient for the compiler to
8196 ``normalize'' the value (convert it to, e.g., 1 or 0) than for the
8197 comparison operators to do so because there may be opportunities to
8198 combine the normalization with other operations.
8201 For equal-length sequences, use a value of 1 or @minus{}1, with @minus{}1 being
8202 slightly preferred on machines with expensive jumps and 1 preferred on
8206 As a second choice, choose a value of @samp{0x80000001} if instructions
8207 exist that set both the sign and low-order bits but do not define the
8211 Otherwise, use a value of @samp{0x80000000}.
8214 Many machines can produce both the value chosen for
8215 @code{STORE_FLAG_VALUE} and its negation in the same number of
8216 instructions. On those machines, you should also define a pattern for
8217 those cases, e.g., one matching
8220 (set @var{A} (neg:@var{m} (ne:@var{m} @var{B} @var{C})))
8223 Some machines can also perform @code{and} or @code{plus} operations on
8224 condition code values with less instructions than the corresponding
8225 @samp{s@var{cond}} insn followed by @code{and} or @code{plus}. On those
8226 machines, define the appropriate patterns. Use the names @code{incscc}
8227 and @code{decscc}, respectively, for the patterns which perform
8228 @code{plus} or @code{minus} operations on condition code values. See
8229 @file{rs6000.md} for some examples. The GNU Superoptizer can be used to
8230 find such instruction sequences on other machines.
8232 You need not define @code{STORE_FLAG_VALUE} if the machine has no store-flag
8235 @findex FLOAT_STORE_FLAG_VALUE
8236 @item FLOAT_STORE_FLAG_VALUE (@var{mode})
8237 A C expression that gives a nonzero @code{REAL_VALUE_TYPE} value that is
8238 returned when comparison operators with floating-point results are true.
8239 Define this macro on machine that have comparison operations that return
8240 floating-point values. If there are no such operations, do not define
8245 An alias for the machine mode for pointers. On most machines, define
8246 this to be the integer mode corresponding to the width of a hardware
8247 pointer; @code{SImode} on 32-bit machine or @code{DImode} on 64-bit machines.
8248 On some machines you must define this to be one of the partial integer
8249 modes, such as @code{PSImode}.
8251 The width of @code{Pmode} must be at least as large as the value of
8252 @code{POINTER_SIZE}. If it is not equal, you must define the macro
8253 @code{POINTERS_EXTEND_UNSIGNED} to specify how pointers are extended
8256 @findex FUNCTION_MODE
8258 An alias for the machine mode used for memory references to functions
8259 being called, in @code{call} RTL expressions. On most machines this
8260 should be @code{QImode}.
8262 @findex INTEGRATE_THRESHOLD
8263 @item INTEGRATE_THRESHOLD (@var{decl})
8264 A C expression for the maximum number of instructions above which the
8265 function @var{decl} should not be inlined. @var{decl} is a
8266 @code{FUNCTION_DECL} node.
8268 The default definition of this macro is 64 plus 8 times the number of
8269 arguments that the function accepts. Some people think a larger
8270 threshold should be used on RISC machines.
8272 @findex STDC_0_IN_SYSTEM_HEADERS
8273 @item STDC_0_IN_SYSTEM_HEADERS
8274 In normal operation, the preprocessor expands @code{__STDC__} to the
8275 constant 1, to signify that GCC conforms to ISO Standard C@. On some
8276 hosts, like Solaris, the system compiler uses a different convention,
8277 where @code{__STDC__} is normally 0, but is 1 if the user specifies
8278 strict conformance to the C Standard.
8280 Defining @code{STDC_0_IN_SYSTEM_HEADERS} makes GNU CPP follows the host
8281 convention when processing system header files, but when processing user
8282 files @code{__STDC__} will always expand to 1.
8284 @findex SCCS_DIRECTIVE
8285 @item SCCS_DIRECTIVE
8286 Define this if the preprocessor should ignore @code{#sccs} directives
8287 and print no error message.
8289 @findex NO_IMPLICIT_EXTERN_C
8290 @item NO_IMPLICIT_EXTERN_C
8291 Define this macro if the system header files support C++ as well as C@.
8292 This macro inhibits the usual method of using system header files in
8293 C++, which is to pretend that the file's contents are enclosed in
8294 @samp{extern "C" @{@dots{}@}}.
8296 @findex HANDLE_PRAGMA
8297 @item HANDLE_PRAGMA (@var{getc}, @var{ungetc}, @var{name})
8298 This macro is no longer supported. You must use
8299 @code{REGISTER_TARGET_PRAGMAS} instead.
8301 @findex REGISTER_TARGET_PRAGMAS
8304 @item REGISTER_TARGET_PRAGMAS (@var{pfile})
8305 Define this macro if you want to implement any target-specific pragmas.
8306 If defined, it is a C expression which makes a series of calls to
8307 @code{cpp_register_pragma} for each pragma, with @var{pfile} passed as
8308 the first argument to to these functions. The macro may also do any
8309 setup required for the pragmas.
8311 The primary reason to define this macro is to provide compatibility with
8312 other compilers for the same target. In general, we discourage
8313 definition of target-specific pragmas for GCC@.
8315 If the pragma can be implemented by attributes then you should consider
8316 defining the target hook @samp{TARGET_INSERT_ATTRIBUTES} as well.
8318 Preprocessor macros that appear on pragma lines are not expanded. All
8319 @samp{#pragma} directives that do not match any registered pragma are
8320 silently ignored, unless the user specifies @option{-Wunknown-pragmas}.
8322 @deftypefun void cpp_register_pragma (cpp_reader *@var{pfile}, const char *@var{space}, const char *@var{name}, void (*@var{callback}) (cpp_reader *))
8324 Each call to @code{cpp_register_pragma} establishes one pragma. The
8325 @var{callback} routine will be called when the preprocessor encounters a
8329 #pragma [@var{space}] @var{name} @dots{}
8332 @var{space} is the case-sensitive namespace of the pragma, or
8333 @code{NULL} to put the pragma in the global namespace. The callback
8334 routine receives @var{pfile} as its first argument, which can be passed
8335 on to cpplib's functions if necessary. You can lex tokens after the
8336 @var{name} by calling @code{c_lex}. Tokens that are not read by the
8337 callback will be silently ignored. The end of the line is indicated by
8338 a token of type @code{CPP_EOF}.
8340 For an example use of this routine, see @file{c4x.h} and the callback
8341 routines defined in @file{c4x-c.c}.
8343 Note that the use of @code{c_lex} is specific to the C and C++
8344 compilers. It will not work in the Java or Fortran compilers, or any
8345 other language compilers for that matter. Thus if @code{c_lex} is going
8346 to be called from target-specific code, it must only be done so when
8347 building the C and C++ compilers. This can be done by defining the
8348 variables @code{c_target_objs} and @code{cxx_target_objs} in the
8349 target entry in the @file{config.gcc} file. These variables should name
8350 the target-specific, language-specific object file which contains the
8351 code that uses @code{c_lex}. Note it will also be necessary to add a
8352 rule to the makefile fragment pointed to by @code{tmake_file} that shows
8353 how to build this object file.
8356 @findex HANDLE_SYSV_PRAGMA
8359 @item HANDLE_SYSV_PRAGMA
8360 Define this macro (to a value of 1) if you want the System V style
8361 pragmas @samp{#pragma pack(<n>)} and @samp{#pragma weak <name>
8362 [=<value>]} to be supported by gcc.
8364 The pack pragma specifies the maximum alignment (in bytes) of fields
8365 within a structure, in much the same way as the @samp{__aligned__} and
8366 @samp{__packed__} @code{__attribute__}s do. A pack value of zero resets
8367 the behavior to the default.
8369 The weak pragma only works if @code{SUPPORTS_WEAK} and
8370 @code{ASM_WEAKEN_LABEL} are defined. If enabled it allows the creation
8371 of specifically named weak labels, optionally with a value.
8373 @findex HANDLE_PRAGMA_PACK_PUSH_POP
8376 @item HANDLE_PRAGMA_PACK_PUSH_POP
8377 Define this macro (to a value of 1) if you want to support the Win32
8378 style pragmas @samp{#pragma pack(push,@var{n})} and @samp{#pragma
8379 pack(pop)}. The @samp{pack(push,@var{n})} pragma specifies the maximum alignment
8380 (in bytes) of fields within a structure, in much the same way as the
8381 @samp{__aligned__} and @samp{__packed__} @code{__attribute__}s do. A
8382 pack value of zero resets the behavior to the default. Successive
8383 invocations of this pragma cause the previous values to be stacked, so
8384 that invocations of @samp{#pragma pack(pop)} will return to the previous
8387 @findex DOLLARS_IN_IDENTIFIERS
8388 @item DOLLARS_IN_IDENTIFIERS
8389 Define this macro to control use of the character @samp{$} in identifier
8390 names. 0 means @samp{$} is not allowed by default; 1 means it is allowed.
8391 1 is the default; there is no need to define this macro in that case.
8392 This macro controls the compiler proper; it does not affect the preprocessor.
8394 @findex NO_DOLLAR_IN_LABEL
8395 @item NO_DOLLAR_IN_LABEL
8396 Define this macro if the assembler does not accept the character
8397 @samp{$} in label names. By default constructors and destructors in
8398 G++ have @samp{$} in the identifiers. If this macro is defined,
8399 @samp{.} is used instead.
8401 @findex NO_DOT_IN_LABEL
8402 @item NO_DOT_IN_LABEL
8403 Define this macro if the assembler does not accept the character
8404 @samp{.} in label names. By default constructors and destructors in G++
8405 have names that use @samp{.}. If this macro is defined, these names
8406 are rewritten to avoid @samp{.}.
8408 @findex DEFAULT_MAIN_RETURN
8409 @item DEFAULT_MAIN_RETURN
8410 Define this macro if the target system expects every program's @code{main}
8411 function to return a standard ``success'' value by default (if no other
8412 value is explicitly returned).
8414 The definition should be a C statement (sans semicolon) to generate the
8415 appropriate rtl instructions. It is used only when compiling the end of
8420 Define this if the target system lacks the function @code{atexit}
8421 from the ISO C standard. If this macro is defined, a default definition
8422 will be provided to support C++. If @code{ON_EXIT} is not defined,
8423 a default @code{exit} function will also be provided.
8427 Define this macro if the target has another way to implement atexit
8428 functionality without replacing @code{exit}. For instance, SunOS 4 has
8429 a similar @code{on_exit} library function.
8431 The definition should be a functional macro which can be used just like
8432 the @code{atexit} function.
8436 Define this if your @code{exit} function needs to do something
8437 besides calling an external function @code{_cleanup} before
8438 terminating with @code{_exit}. The @code{EXIT_BODY} macro is
8439 only needed if @code{NEED_ATEXIT} is defined and @code{ON_EXIT} is not
8442 @findex INSN_SETS_ARE_DELAYED
8443 @item INSN_SETS_ARE_DELAYED (@var{insn})
8444 Define this macro as a C expression that is nonzero if it is safe for the
8445 delay slot scheduler to place instructions in the delay slot of @var{insn},
8446 even if they appear to use a resource set or clobbered in @var{insn}.
8447 @var{insn} is always a @code{jump_insn} or an @code{insn}; GCC knows that
8448 every @code{call_insn} has this behavior. On machines where some @code{insn}
8449 or @code{jump_insn} is really a function call and hence has this behavior,
8450 you should define this macro.
8452 You need not define this macro if it would always return zero.
8454 @findex INSN_REFERENCES_ARE_DELAYED
8455 @item INSN_REFERENCES_ARE_DELAYED (@var{insn})
8456 Define this macro as a C expression that is nonzero if it is safe for the
8457 delay slot scheduler to place instructions in the delay slot of @var{insn},
8458 even if they appear to set or clobber a resource referenced in @var{insn}.
8459 @var{insn} is always a @code{jump_insn} or an @code{insn}. On machines where
8460 some @code{insn} or @code{jump_insn} is really a function call and its operands
8461 are registers whose use is actually in the subroutine it calls, you should
8462 define this macro. Doing so allows the delay slot scheduler to move
8463 instructions which copy arguments into the argument registers into the delay
8466 You need not define this macro if it would always return zero.
8468 @findex MACHINE_DEPENDENT_REORG
8469 @item MACHINE_DEPENDENT_REORG (@var{insn})
8470 In rare cases, correct code generation requires extra machine
8471 dependent processing between the second jump optimization pass and
8472 delayed branch scheduling. On those machines, define this macro as a C
8473 statement to act on the code starting at @var{insn}.
8475 @findex MULTIPLE_SYMBOL_SPACES
8476 @item MULTIPLE_SYMBOL_SPACES
8477 Define this macro if in some cases global symbols from one translation
8478 unit may not be bound to undefined symbols in another translation unit
8479 without user intervention. For instance, under Microsoft Windows
8480 symbols must be explicitly imported from shared libraries (DLLs).
8482 @findex MD_ASM_CLOBBERS
8483 @item MD_ASM_CLOBBERS (@var{clobbers})
8484 A C statement that adds to @var{clobbers} @code{STRING_CST} trees for
8485 any hard regs the port wishes to automatically clobber for all asms.
8487 @findex MAX_INTEGER_COMPUTATION_MODE
8488 @item MAX_INTEGER_COMPUTATION_MODE
8489 Define this to the largest integer machine mode which can be used for
8490 operations other than load, store and copy operations.
8492 You need only define this macro if the target holds values larger than
8493 @code{word_mode} in general purpose registers. Most targets should not define
8496 @findex MATH_LIBRARY
8498 Define this macro as a C string constant for the linker argument to link
8499 in the system math library, or @samp{""} if the target does not have a
8500 separate math library.
8502 You need only define this macro if the default of @samp{"-lm"} is wrong.
8504 @findex LIBRARY_PATH_ENV
8505 @item LIBRARY_PATH_ENV
8506 Define this macro as a C string constant for the environment variable that
8507 specifies where the linker should look for libraries.
8509 You need only define this macro if the default of @samp{"LIBRARY_PATH"}
8512 @findex TARGET_HAS_F_SETLKW
8513 @item TARGET_HAS_F_SETLKW
8514 Define this macro if the target supports file locking with fcntl / F_SETLKW@.
8515 Note that this functionality is part of POSIX@.
8516 Defining @code{TARGET_HAS_F_SETLKW} will enable the test coverage code
8517 to use file locking when exiting a program, which avoids race conditions
8518 if the program has forked.
8520 @findex MAX_CONDITIONAL_EXECUTE
8521 @item MAX_CONDITIONAL_EXECUTE
8523 A C expression for the maximum number of instructions to execute via
8524 conditional execution instructions instead of a branch. A value of
8525 @code{BRANCH_COST}+1 is the default if the machine does not use cc0, and
8526 1 if it does use cc0.
8528 @findex IFCVT_MODIFY_TESTS
8529 @item IFCVT_MODIFY_TESTS
8530 A C expression to modify the tests in @code{TRUE_EXPR}, and
8531 @code{FALSE_EXPR} for use in converting insns in @code{TEST_BB},
8532 @code{THEN_BB}, @code{ELSE_BB}, and @code{JOIN_BB} basic blocks to
8533 conditional execution. Set either @code{TRUE_EXPR} or @code{FALSE_EXPR}
8534 to a null pointer if the tests cannot be converted.
8536 @findex IFCVT_MODIFY_INSN
8537 @item IFCVT_MODIFY_INSN
8538 A C expression to modify the @code{PATTERN} of an @code{INSN} that is to
8539 be converted to conditional execution format.
8541 @findex IFCVT_MODIFY_FINAL
8542 @item IFCVT_MODIFY_FINAL
8543 A C expression to perform any final machine dependent modifications in
8544 converting code to conditional execution in the basic blocks
8545 @code{TEST_BB}, @code{THEN_BB}, @code{ELSE_BB}, and @code{JOIN_BB}.
8547 @findex IFCVT_MODIFY_CANCEL
8548 @item IFCVT_MODIFY_CANCEL
8549 A C expression to cancel any machine dependent modifications in
8550 converting code to conditional execution in the basic blocks
8551 @code{TEST_BB}, @code{THEN_BB}, @code{ELSE_BB}, and @code{JOIN_BB}.
8554 @deftypefn {Target Hook} void TARGET_INIT_BUILTINS ()
8555 Define this hook if you have any machine-specific built-in functions
8556 that need to be defined. It should be a function that performs the
8559 Machine specific built-in functions can be useful to expand special machine
8560 instructions that would otherwise not normally be generated because
8561 they have no equivalent in the source language (for example, SIMD vector
8562 instructions or prefetch instructions).
8564 To create a built-in function, call the function @code{builtin_function}
8565 which is defined by the language front end. You can use any type nodes set
8566 up by @code{build_common_tree_nodes} and @code{build_common_tree_nodes_2};
8567 only language front ends that use those two functions will call
8568 @samp{TARGET_INIT_BUILTINS}.
8571 @deftypefn {Target Hook} rtx TARGET_EXPAND_BUILTIN (tree @var{exp}, rtx @var{target}, rtx @var{subtarget}, enum machine_mode @var{mode}, int @var{ignore})
8573 Expand a call to a machine specific built-in function that was set up by
8574 @samp{TARGET_INIT_BUILTINS}. @var{exp} is the expression for the
8575 function call; the result should go to @var{target} if that is
8576 convenient, and have mode @var{mode} if that is convenient.
8577 @var{subtarget} may be used as the target for computing one of
8578 @var{exp}'s operands. @var{ignore} is nonzero if the value is to be
8579 ignored. This function should return the result of the call to the
8584 @findex MD_CAN_REDIRECT_BRANCH
8585 @item MD_CAN_REDIRECT_BRANCH(@var{branch1}, @var{branch2})
8587 Take a branch insn in @var{branch1} and another in @var{branch2}.
8588 Return true if redirecting @var{branch1} to the destination of
8589 @var{branch2} is possible.
8591 On some targets, branches may have a limited range. Optimizing the
8592 filling of delay slots can result in branches being redirected, and this
8593 may in turn cause a branch offset to overflow.
8595 @findex ALLOCATE_INITIAL_VALUE
8596 @item ALLOCATE_INITIAL_VALUE(@var{hard_reg})
8598 When the initial value of a hard register has been copied in a pseudo
8599 register, it is often not necessary to actually allocate another register
8600 to this pseudo register, because the original hard register or a stack slot
8601 it has been saved into can be used. @code{ALLOCATE_INITIAL_VALUE}, if
8602 defined, is called at the start of register allocation once for each
8603 hard register that had its initial value copied by using
8604 @code{get_func_hard_reg_initial_val} or @code{get_hard_reg_initial_val}.
8605 Possible values are @code{NULL_RTX}, if you don't want
8606 to do any special allocation, a @code{REG} rtx---that would typically be
8607 the hard register itself, if it is known not to be clobbered---or a
8609 If you are returning a @code{MEM}, this is only a hint for the allocator;
8610 it might decide to use another register anyways.
8611 You may use @code{current_function_leaf_function} in the definition of the
8612 macro, functions that use @code{REG_N_SETS}, to determine if the hard
8613 register in question will not be clobbered.