1 @c Copyright (C) 1988,1989,1992,1993,1994,1995,1996,1997,1998,1999,2000,2001,
2 @c 2002, 2003, 2004, 2005 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 * Registers:: Naming and describing the hardware registers.
35 * Register Classes:: Defining the classes of hardware registers.
36 * Stack and Calling:: Defining which way the stack grows and by how much.
37 * Varargs:: Defining the varargs macros.
38 * Trampolines:: Code set up at run time to enter a nested function.
39 * Library Calls:: Controlling how library routines are implicitly called.
40 * Addressing Modes:: Defining addressing modes valid for memory operands.
41 * Condition Code:: Defining how insns update the condition code.
42 * Costs:: Defining relative costs of different operations.
43 * Scheduling:: Adjusting the behavior of the instruction scheduler.
44 * Sections:: Dividing storage into text, data, and other sections.
45 * PIC:: Macros for position independent code.
46 * Assembler Format:: Defining how to write insns and pseudo-ops to output.
47 * Debugging Info:: Defining the format of debugging output.
48 * Floating Point:: Handling floating point for cross-compilers.
49 * Mode Switching:: Insertion of mode-switching instructions.
50 * Target Attributes:: Defining target-specific uses of @code{__attribute__}.
51 * MIPS Coprocessors:: MIPS coprocessor support and how to customize it.
52 * PCH Target:: Validity checking for precompiled headers.
53 * C++ ABI:: Controlling C++ ABI changes.
54 * Misc:: Everything else.
57 @node Target Structure
58 @section The Global @code{targetm} Variable
60 @cindex target functions
62 @deftypevar {struct gcc_target} targetm
63 The target @file{.c} file must define the global @code{targetm} variable
64 which contains pointers to functions and data relating to the target
65 machine. The variable is declared in @file{target.h};
66 @file{target-def.h} defines the macro @code{TARGET_INITIALIZER} which is
67 used to initialize the variable, and macros for the default initializers
68 for elements of the structure. The @file{.c} file should override those
69 macros for which the default definition is inappropriate. For example:
72 #include "target-def.h"
74 /* @r{Initialize the GCC target structure.} */
76 #undef TARGET_COMP_TYPE_ATTRIBUTES
77 #define TARGET_COMP_TYPE_ATTRIBUTES @var{machine}_comp_type_attributes
79 struct gcc_target targetm = TARGET_INITIALIZER;
83 Where a macro should be defined in the @file{.c} file in this manner to
84 form part of the @code{targetm} structure, it is documented below as a
85 ``Target Hook'' with a prototype. Many macros will change in future
86 from being defined in the @file{.h} file to being part of the
87 @code{targetm} structure.
90 @section Controlling the Compilation Driver, @file{gcc}
92 @cindex controlling the compilation driver
94 @c prevent bad page break with this line
95 You can control the compilation driver.
97 @defmac 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
110 @defmac 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
124 @defmac 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.
139 @defmac 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{""}.
146 @defmac 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 @option{-malt-abi},
157 @option{-EB}, and @option{-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" @}
166 @defmac DRIVER_SELF_SPECS
167 A list of specs for the driver itself. It should be a suitable
168 initializer for an array of strings, with no surrounding braces.
170 The driver applies these specs to its own command line between loading
171 default @file{specs} files (but not command-line specified ones) and
172 choosing the multilib directory or running any subcommands. It
173 applies them in the order given, so each spec can depend on the
174 options added by earlier ones. It is also possible to remove options
175 using @samp{%<@var{option}} in the usual way.
177 This macro can be useful when a port has several interdependent target
178 options. It provides a way of standardizing the command line so
179 that the other specs are easier to write.
181 Do not define this macro if it does not need to do anything.
184 @defmac OPTION_DEFAULT_SPECS
185 A list of specs used to support configure-time default options (i.e.@:
186 @option{--with} options) in the driver. It should be a suitable initializer
187 for an array of structures, each containing two strings, without the
188 outermost pair of surrounding braces.
190 The first item in the pair is the name of the default. This must match
191 the code in @file{config.gcc} for the target. The second item is a spec
192 to apply if a default with this name was specified. The string
193 @samp{%(VALUE)} in the spec will be replaced by the value of the default
194 everywhere it occurs.
196 The driver will apply these specs to its own command line between loading
197 default @file{specs} files and processing @code{DRIVER_SELF_SPECS}, using
198 the same mechanism as @code{DRIVER_SELF_SPECS}.
200 Do not define this macro if it does not need to do anything.
204 A C string constant that tells the GCC driver program options to
205 pass to CPP@. It can also specify how to translate options you
206 give to GCC into options for GCC to pass to the CPP@.
208 Do not define this macro if it does not need to do anything.
211 @defmac CPLUSPLUS_CPP_SPEC
212 This macro is just like @code{CPP_SPEC}, but is used for C++, rather
213 than C@. If you do not define this macro, then the value of
214 @code{CPP_SPEC} (if any) will be used instead.
218 A C string constant that tells the GCC driver program options to
219 pass to @code{cc1}, @code{cc1plus}, @code{f771}, and the other language
221 It can also specify how to translate options you give to GCC into options
222 for GCC to pass to front ends.
224 Do not define this macro if it does not need to do anything.
228 A C string constant that tells the GCC driver program options to
229 pass to @code{cc1plus}. It can also specify how to translate options you
230 give to GCC into options for GCC to pass to the @code{cc1plus}.
232 Do not define this macro if it does not need to do anything.
233 Note that everything defined in CC1_SPEC is already passed to
234 @code{cc1plus} so there is no need to duplicate the contents of
235 CC1_SPEC in CC1PLUS_SPEC@.
239 A C string constant that tells the GCC driver program options to
240 pass to the assembler. It can also specify how to translate options
241 you give to GCC into options for GCC to pass to the assembler.
242 See the file @file{sun3.h} for an example of this.
244 Do not define this macro if it does not need to do anything.
247 @defmac ASM_FINAL_SPEC
248 A C string constant that tells the GCC driver program how to
249 run any programs which cleanup after the normal assembler.
250 Normally, this is not needed. See the file @file{mips.h} for
253 Do not define this macro if it does not need to do anything.
256 @defmac AS_NEEDS_DASH_FOR_PIPED_INPUT
257 Define this macro, with no value, if the driver should give the assembler
258 an argument consisting of a single dash, @option{-}, to instruct it to
259 read from its standard input (which will be a pipe connected to the
260 output of the compiler proper). This argument is given after any
261 @option{-o} option specifying the name of the output file.
263 If you do not define this macro, the assembler is assumed to read its
264 standard input if given no non-option arguments. If your assembler
265 cannot read standard input at all, use a @samp{%@{pipe:%e@}} construct;
266 see @file{mips.h} for instance.
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.
296 @defmac REAL_LIBGCC_SPEC
297 By default, if @code{ENABLE_SHARED_LIBGCC} is defined, the
298 @code{LIBGCC_SPEC} is not directly used by the driver program but is
299 instead modified to refer to different versions of @file{libgcc.a}
300 depending on the values of the command line flags @option{-static},
301 @option{-shared}, @option{-static-libgcc}, and @option{-shared-libgcc}. On
302 targets where these modifications are inappropriate, define
303 @code{REAL_LIBGCC_SPEC} instead. @code{REAL_LIBGCC_SPEC} tells the
304 driver how to place a reference to @file{libgcc} on the link command
305 line, but, unlike @code{LIBGCC_SPEC}, it is used unmodified.
308 @defmac USE_LD_AS_NEEDED
309 A macro that controls the modifications to @code{LIBGCC_SPEC}
310 mentioned in @code{REAL_LIBGCC_SPEC}. If nonzero, a spec will be
311 generated that uses --as-needed and the shared libgcc in place of the
312 static exception handler library, when linking without any of
313 @code{-static}, @code{-static-libgcc}, or @code{-shared-libgcc}.
317 If defined, this C string constant is added to @code{LINK_SPEC}.
318 When @code{USE_LD_AS_NEEDED} is zero or undefined, it also affects
319 the modifications to @code{LIBGCC_SPEC} mentioned in
320 @code{REAL_LIBGCC_SPEC}.
323 @defmac STARTFILE_SPEC
324 Another C string constant used much like @code{LINK_SPEC}. The
325 difference between the two is that @code{STARTFILE_SPEC} is used at
326 the very beginning of the command given to the linker.
328 If this macro is not defined, a default is provided that loads the
329 standard C startup file from the usual place. See @file{gcc.c}.
333 Another C string constant used much like @code{LINK_SPEC}. The
334 difference between the two is that @code{ENDFILE_SPEC} is used at
335 the very end of the command given to the linker.
337 Do not define this macro if it does not need to do anything.
340 @defmac THREAD_MODEL_SPEC
341 GCC @code{-v} will print the thread model GCC was configured to use.
342 However, this doesn't work on platforms that are multilibbed on thread
343 models, such as AIX 4.3. On such platforms, define
344 @code{THREAD_MODEL_SPEC} such that it evaluates to a string without
345 blanks that names one of the recognized thread models. @code{%*}, the
346 default value of this macro, will expand to the value of
347 @code{thread_file} set in @file{config.gcc}.
350 @defmac SYSROOT_SUFFIX_SPEC
351 Define this macro to add a suffix to the target sysroot when GCC is
352 configured with a sysroot. This will cause GCC to search for usr/lib,
353 et al, within sysroot+suffix.
356 @defmac SYSROOT_HEADERS_SUFFIX_SPEC
357 Define this macro to add a headers_suffix to the target sysroot when
358 GCC is configured with a sysroot. This will cause GCC to pass the
359 updated sysroot+headers_suffix to CPP, causing it to search for
360 usr/include, et al, within sysroot+headers_suffix.
364 Define this macro to provide additional specifications to put in the
365 @file{specs} file that can be used in various specifications like
368 The definition should be an initializer for an array of structures,
369 containing a string constant, that defines the specification name, and a
370 string constant that provides the specification.
372 Do not define this macro if it does not need to do anything.
374 @code{EXTRA_SPECS} is useful when an architecture contains several
375 related targets, which have various @code{@dots{}_SPECS} which are similar
376 to each other, and the maintainer would like one central place to keep
379 For example, the PowerPC System V.4 targets use @code{EXTRA_SPECS} to
380 define either @code{_CALL_SYSV} when the System V calling sequence is
381 used or @code{_CALL_AIX} when the older AIX-based calling sequence is
384 The @file{config/rs6000/rs6000.h} target file defines:
387 #define EXTRA_SPECS \
388 @{ "cpp_sysv_default", CPP_SYSV_DEFAULT @},
390 #define CPP_SYS_DEFAULT ""
393 The @file{config/rs6000/sysv.h} target file defines:
397 "%@{posix: -D_POSIX_SOURCE @} \
398 %@{mcall-sysv: -D_CALL_SYSV @} \
399 %@{!mcall-sysv: %(cpp_sysv_default) @} \
400 %@{msoft-float: -D_SOFT_FLOAT@} %@{mcpu=403: -D_SOFT_FLOAT@}"
402 #undef CPP_SYSV_DEFAULT
403 #define CPP_SYSV_DEFAULT "-D_CALL_SYSV"
406 while the @file{config/rs6000/eabiaix.h} target file defines
407 @code{CPP_SYSV_DEFAULT} as:
410 #undef CPP_SYSV_DEFAULT
411 #define CPP_SYSV_DEFAULT "-D_CALL_AIX"
415 @defmac LINK_LIBGCC_SPECIAL_1
416 Define this macro if the driver program should find the library
417 @file{libgcc.a}. If you do not define this macro, the driver program will pass
418 the argument @option{-lgcc} to tell the linker to do the search.
421 @defmac LINK_GCC_C_SEQUENCE_SPEC
422 The sequence in which libgcc and libc are specified to the linker.
423 By default this is @code{%G %L %G}.
426 @defmac LINK_COMMAND_SPEC
427 A C string constant giving the complete command line need to execute the
428 linker. When you do this, you will need to update your port each time a
429 change is made to the link command line within @file{gcc.c}. Therefore,
430 define this macro only if you need to completely redefine the command
431 line for invoking the linker and there is no other way to accomplish
432 the effect you need. Overriding this macro may be avoidable by overriding
433 @code{LINK_GCC_C_SEQUENCE_SPEC} instead.
436 @defmac LINK_ELIMINATE_DUPLICATE_LDIRECTORIES
437 A nonzero value causes @command{collect2} to remove duplicate @option{-L@var{directory}} search
438 directories from linking commands. Do not give it a nonzero value if
439 removing duplicate search directories changes the linker's semantics.
442 @defmac MULTILIB_DEFAULTS
443 Define this macro as a C expression for the initializer of an array of
444 string to tell the driver program which options are defaults for this
445 target and thus do not need to be handled specially when using
446 @code{MULTILIB_OPTIONS}.
448 Do not define this macro if @code{MULTILIB_OPTIONS} is not defined in
449 the target makefile fragment or if none of the options listed in
450 @code{MULTILIB_OPTIONS} are set by default.
451 @xref{Target Fragment}.
454 @defmac RELATIVE_PREFIX_NOT_LINKDIR
455 Define this macro to tell @command{gcc} that it should only translate
456 a @option{-B} prefix into a @option{-L} linker option if the prefix
457 indicates an absolute file name.
460 @defmac MD_EXEC_PREFIX
461 If defined, this macro is an additional prefix to try after
462 @code{STANDARD_EXEC_PREFIX}. @code{MD_EXEC_PREFIX} is not searched
463 when the @option{-b} option is used, or the compiler is built as a cross
464 compiler. If you define @code{MD_EXEC_PREFIX}, then be sure to add it
465 to the list of directories used to find the assembler in @file{configure.in}.
468 @defmac STANDARD_STARTFILE_PREFIX
469 Define this macro as a C string constant if you wish to override the
470 standard choice of @code{libdir} as the default prefix to
471 try when searching for startup files such as @file{crt0.o}.
472 @code{STANDARD_STARTFILE_PREFIX} is not searched when the compiler
473 is built as a cross compiler.
476 @defmac STANDARD_STARTFILE_PREFIX_1
477 Define this macro as a C string constant if you wish to override the
478 standard choice of @code{/lib} as a prefix to try after the default prefix
479 when searching for startup files such as @file{crt0.o}.
480 @code{STANDARD_STARTFILE_PREFIX_1} is not searched when the compiler
481 is built as a cross compiler.
484 @defmac STANDARD_STARTFILE_PREFIX_2
485 Define this macro as a C string constant if you wish to override the
486 standard choice of @code{/lib} as yet another prefix to try after the
487 default prefix when searching for startup files such as @file{crt0.o}.
488 @code{STANDARD_STARTFILE_PREFIX_2} is not searched when the compiler
489 is built as a cross compiler.
492 @defmac MD_STARTFILE_PREFIX
493 If defined, this macro supplies an additional prefix to try after the
494 standard prefixes. @code{MD_EXEC_PREFIX} is not searched when the
495 @option{-b} option is used, or when the compiler is built as a cross
499 @defmac MD_STARTFILE_PREFIX_1
500 If defined, this macro supplies yet another prefix to try after the
501 standard prefixes. It is not searched when the @option{-b} option is
502 used, or when the compiler is built as a cross compiler.
505 @defmac INIT_ENVIRONMENT
506 Define this macro as a C string constant if you wish to set environment
507 variables for programs called by the driver, such as the assembler and
508 loader. The driver passes the value of this macro to @code{putenv} to
509 initialize the necessary environment variables.
512 @defmac LOCAL_INCLUDE_DIR
513 Define this macro as a C string constant if you wish to override the
514 standard choice of @file{/usr/local/include} as the default prefix to
515 try when searching for local header files. @code{LOCAL_INCLUDE_DIR}
516 comes before @code{SYSTEM_INCLUDE_DIR} in the search order.
518 Cross compilers do not search either @file{/usr/local/include} or its
522 @defmac MODIFY_TARGET_NAME
523 Define this macro if you wish to define command-line switches that
524 modify the default target name.
526 For each switch, you can include a string to be appended to the first
527 part of the configuration name or a string to be deleted from the
528 configuration name, if present. The definition should be an initializer
529 for an array of structures. Each array element should have three
530 elements: the switch name (a string constant, including the initial
531 dash), one of the enumeration codes @code{ADD} or @code{DELETE} to
532 indicate whether the string should be inserted or deleted, and the string
533 to be inserted or deleted (a string constant).
535 For example, on a machine where @samp{64} at the end of the
536 configuration name denotes a 64-bit target and you want the @option{-32}
537 and @option{-64} switches to select between 32- and 64-bit targets, you would
541 #define MODIFY_TARGET_NAME \
542 @{ @{ "-32", DELETE, "64"@}, \
543 @{"-64", ADD, "64"@}@}
547 @defmac SYSTEM_INCLUDE_DIR
548 Define this macro as a C string constant if you wish to specify a
549 system-specific directory to search for header files before the standard
550 directory. @code{SYSTEM_INCLUDE_DIR} comes before
551 @code{STANDARD_INCLUDE_DIR} in the search order.
553 Cross compilers do not use this macro and do not search the directory
557 @defmac STANDARD_INCLUDE_DIR
558 Define this macro as a C string constant if you wish to override the
559 standard choice of @file{/usr/include} as the default prefix to
560 try when searching for header files.
562 Cross compilers ignore this macro and do not search either
563 @file{/usr/include} or its replacement.
566 @defmac STANDARD_INCLUDE_COMPONENT
567 The ``component'' corresponding to @code{STANDARD_INCLUDE_DIR}.
568 See @code{INCLUDE_DEFAULTS}, below, for the description of components.
569 If you do not define this macro, no component is used.
572 @defmac INCLUDE_DEFAULTS
573 Define this macro if you wish to override the entire default search path
574 for include files. For a native compiler, the default search path
575 usually consists of @code{GCC_INCLUDE_DIR}, @code{LOCAL_INCLUDE_DIR},
576 @code{SYSTEM_INCLUDE_DIR}, @code{GPLUSPLUS_INCLUDE_DIR}, and
577 @code{STANDARD_INCLUDE_DIR}. In addition, @code{GPLUSPLUS_INCLUDE_DIR}
578 and @code{GCC_INCLUDE_DIR} are defined automatically by @file{Makefile},
579 and specify private search areas for GCC@. The directory
580 @code{GPLUSPLUS_INCLUDE_DIR} is used only for C++ programs.
582 The definition should be an initializer for an array of structures.
583 Each array element should have four elements: the directory name (a
584 string constant), the component name (also a string constant), a flag
585 for C++-only directories,
586 and a flag showing that the includes in the directory don't need to be
587 wrapped in @code{extern @samp{C}} when compiling C++. Mark the end of
588 the array with a null element.
590 The component name denotes what GNU package the include file is part of,
591 if any, in all uppercase letters. For example, it might be @samp{GCC}
592 or @samp{BINUTILS}. If the package is part of a vendor-supplied
593 operating system, code the component name as @samp{0}.
595 For example, here is the definition used for VAX/VMS:
598 #define INCLUDE_DEFAULTS \
600 @{ "GNU_GXX_INCLUDE:", "G++", 1, 1@}, \
601 @{ "GNU_CC_INCLUDE:", "GCC", 0, 0@}, \
602 @{ "SYS$SYSROOT:[SYSLIB.]", 0, 0, 0@}, \
609 Here is the order of prefixes tried for exec files:
613 Any prefixes specified by the user with @option{-B}.
616 The environment variable @code{GCC_EXEC_PREFIX}, if any.
619 The directories specified by the environment variable @code{COMPILER_PATH}.
622 The macro @code{STANDARD_EXEC_PREFIX}.
625 @file{/usr/lib/gcc/}.
628 The macro @code{MD_EXEC_PREFIX}, if any.
631 Here is the order of prefixes tried for startfiles:
635 Any prefixes specified by the user with @option{-B}.
638 The environment variable @code{GCC_EXEC_PREFIX}, if any.
641 The directories specified by the environment variable @code{LIBRARY_PATH}
642 (or port-specific name; native only, cross compilers do not use this).
645 The macro @code{STANDARD_EXEC_PREFIX}.
648 @file{/usr/lib/gcc/}.
651 The macro @code{MD_EXEC_PREFIX}, if any.
654 The macro @code{MD_STARTFILE_PREFIX}, if any.
657 The macro @code{STANDARD_STARTFILE_PREFIX}.
666 @node Run-time Target
667 @section Run-time Target Specification
668 @cindex run-time target specification
669 @cindex predefined macros
670 @cindex target specifications
672 @c prevent bad page break with this line
673 Here are run-time target specifications.
675 @defmac TARGET_CPU_CPP_BUILTINS ()
676 This function-like macro expands to a block of code that defines
677 built-in preprocessor macros and assertions for the target cpu, using
678 the functions @code{builtin_define}, @code{builtin_define_std} and
679 @code{builtin_assert}. When the front end
680 calls this macro it provides a trailing semicolon, and since it has
681 finished command line option processing your code can use those
684 @code{builtin_assert} takes a string in the form you pass to the
685 command-line option @option{-A}, such as @code{cpu=mips}, and creates
686 the assertion. @code{builtin_define} takes a string in the form
687 accepted by option @option{-D} and unconditionally defines the macro.
689 @code{builtin_define_std} takes a string representing the name of an
690 object-like macro. If it doesn't lie in the user's namespace,
691 @code{builtin_define_std} defines it unconditionally. Otherwise, it
692 defines a version with two leading underscores, and another version
693 with two leading and trailing underscores, and defines the original
694 only if an ISO standard was not requested on the command line. For
695 example, passing @code{unix} defines @code{__unix}, @code{__unix__}
696 and possibly @code{unix}; passing @code{_mips} defines @code{__mips},
697 @code{__mips__} and possibly @code{_mips}, and passing @code{_ABI64}
698 defines only @code{_ABI64}.
700 You can also test for the C dialect being compiled. The variable
701 @code{c_language} is set to one of @code{clk_c}, @code{clk_cplusplus}
702 or @code{clk_objective_c}. Note that if we are preprocessing
703 assembler, this variable will be @code{clk_c} but the function-like
704 macro @code{preprocessing_asm_p()} will return true, so you might want
705 to check for that first. If you need to check for strict ANSI, the
706 variable @code{flag_iso} can be used. The function-like macro
707 @code{preprocessing_trad_p()} can be used to check for traditional
711 @defmac TARGET_OS_CPP_BUILTINS ()
712 Similarly to @code{TARGET_CPU_CPP_BUILTINS} but this macro is optional
713 and is used for the target operating system instead.
716 @defmac TARGET_OBJFMT_CPP_BUILTINS ()
717 Similarly to @code{TARGET_CPU_CPP_BUILTINS} but this macro is optional
718 and is used for the target object format. @file{elfos.h} uses this
719 macro to define @code{__ELF__}, so you probably do not need to define
723 @deftypevar {extern int} target_flags
724 This variable is declared in @file{options.h}, which is included before
725 any target-specific headers.
728 @deftypevar {Target Hook} int TARGET_DEFAULT_TARGET_FLAGS
729 This variable specifies the initial value of @code{target_flags}.
730 Its default setting is 0.
733 @cindex optional hardware or system features
734 @cindex features, optional, in system conventions
736 @deftypefn {Target Hook} bool TARGET_HANDLE_OPTION (size_t @var{code}, const char *@var{arg}, int @var{value})
737 This hook is called whenever the user specifies one of the
738 target-specific options described by the @file{.opt} definition files
739 (@pxref{Options}). It has the opportunity to do some option-specific
740 processing and should return true if the option is valid. The default
741 definition does nothing but return true.
743 @var{code} specifies the @code{OPT_@var{name}} enumeration value
744 associated with the selected option; @var{name} is just a rendering of
745 the option name in which non-alphanumeric characters are replaced by
746 underscores. @var{arg} specifies the string argument and is null if
747 no argument was given. If the option is flagged as a @code{UInteger}
748 (@pxref{Option properties}), @var{value} is the numeric value of the
749 argument. Otherwise @var{value} is 1 if the positive form of the
750 option was used and 0 if the ``no-'' form was.
753 @defmac TARGET_VERSION
754 This macro is a C statement to print on @code{stderr} a string
755 describing the particular machine description choice. Every machine
756 description should define @code{TARGET_VERSION}. For example:
760 #define TARGET_VERSION \
761 fprintf (stderr, " (68k, Motorola syntax)");
763 #define TARGET_VERSION \
764 fprintf (stderr, " (68k, MIT syntax)");
769 @defmac OVERRIDE_OPTIONS
770 Sometimes certain combinations of command options do not make sense on
771 a particular target machine. You can define a macro
772 @code{OVERRIDE_OPTIONS} to take account of this. This macro, if
773 defined, is executed once just after all the command options have been
776 Don't use this macro to turn on various extra optimizations for
777 @option{-O}. That is what @code{OPTIMIZATION_OPTIONS} is for.
780 @defmac OPTIMIZATION_OPTIONS (@var{level}, @var{size})
781 Some machines may desire to change what optimizations are performed for
782 various optimization levels. This macro, if defined, is executed once
783 just after the optimization level is determined and before the remainder
784 of the command options have been parsed. Values set in this macro are
785 used as the default values for the other command line options.
787 @var{level} is the optimization level specified; 2 if @option{-O2} is
788 specified, 1 if @option{-O} is specified, and 0 if neither is specified.
790 @var{size} is nonzero if @option{-Os} is specified and zero otherwise.
792 You should not use this macro to change options that are not
793 machine-specific. These should uniformly selected by the same
794 optimization level on all supported machines. Use this macro to enable
795 machine-specific optimizations.
797 @strong{Do not examine @code{write_symbols} in
798 this macro!} The debugging options are not supposed to alter the
802 @defmac CAN_DEBUG_WITHOUT_FP
803 Define this macro if debugging can be performed even without a frame
804 pointer. If this macro is defined, GCC will turn on the
805 @option{-fomit-frame-pointer} option whenever @option{-O} is specified.
808 @node Per-Function Data
809 @section Defining data structures for per-function information.
810 @cindex per-function data
811 @cindex data structures
813 If the target needs to store information on a per-function basis, GCC
814 provides a macro and a couple of variables to allow this. Note, just
815 using statics to store the information is a bad idea, since GCC supports
816 nested functions, so you can be halfway through encoding one function
817 when another one comes along.
819 GCC defines a data structure called @code{struct function} which
820 contains all of the data specific to an individual function. This
821 structure contains a field called @code{machine} whose type is
822 @code{struct machine_function *}, which can be used by targets to point
823 to their own specific data.
825 If a target needs per-function specific data it should define the type
826 @code{struct machine_function} and also the macro @code{INIT_EXPANDERS}.
827 This macro should be used to initialize the function pointer
828 @code{init_machine_status}. This pointer is explained below.
830 One typical use of per-function, target specific data is to create an
831 RTX to hold the register containing the function's return address. This
832 RTX can then be used to implement the @code{__builtin_return_address}
833 function, for level 0.
835 Note---earlier implementations of GCC used a single data area to hold
836 all of the per-function information. Thus when processing of a nested
837 function began the old per-function data had to be pushed onto a
838 stack, and when the processing was finished, it had to be popped off the
839 stack. GCC used to provide function pointers called
840 @code{save_machine_status} and @code{restore_machine_status} to handle
841 the saving and restoring of the target specific information. Since the
842 single data area approach is no longer used, these pointers are no
845 @defmac INIT_EXPANDERS
846 Macro called to initialize any target specific information. This macro
847 is called once per function, before generation of any RTL has begun.
848 The intention of this macro is to allow the initialization of the
849 function pointer @code{init_machine_status}.
852 @deftypevar {void (*)(struct function *)} init_machine_status
853 If this function pointer is non-@code{NULL} it will be called once per
854 function, before function compilation starts, in order to allow the
855 target to perform any target specific initialization of the
856 @code{struct function} structure. It is intended that this would be
857 used to initialize the @code{machine} of that structure.
859 @code{struct machine_function} structures are expected to be freed by GC@.
860 Generally, any memory that they reference must be allocated by using
861 @code{ggc_alloc}, including the structure itself.
865 @section Storage Layout
866 @cindex storage layout
868 Note that the definitions of the macros in this table which are sizes or
869 alignments measured in bits do not need to be constant. They can be C
870 expressions that refer to static variables, such as the @code{target_flags}.
871 @xref{Run-time Target}.
873 @defmac BITS_BIG_ENDIAN
874 Define this macro to have the value 1 if the most significant bit in a
875 byte has the lowest number; otherwise define it to have the value zero.
876 This means that bit-field instructions count from the most significant
877 bit. If the machine has no bit-field instructions, then this must still
878 be defined, but it doesn't matter which value it is defined to. This
879 macro need not be a constant.
881 This macro does not affect the way structure fields are packed into
882 bytes or words; that is controlled by @code{BYTES_BIG_ENDIAN}.
885 @defmac BYTES_BIG_ENDIAN
886 Define this macro to have the value 1 if the most significant byte in a
887 word has the lowest number. This macro need not be a constant.
890 @defmac WORDS_BIG_ENDIAN
891 Define this macro to have the value 1 if, in a multiword object, the
892 most significant word has the lowest number. This applies to both
893 memory locations and registers; GCC fundamentally assumes that the
894 order of words in memory is the same as the order in registers. This
895 macro need not be a constant.
898 @defmac LIBGCC2_WORDS_BIG_ENDIAN
899 Define this macro if @code{WORDS_BIG_ENDIAN} is not constant. This must be a
900 constant value with the same meaning as @code{WORDS_BIG_ENDIAN}, which will be
901 used only when compiling @file{libgcc2.c}. Typically the value will be set
902 based on preprocessor defines.
905 @defmac FLOAT_WORDS_BIG_ENDIAN
906 Define this macro to have the value 1 if @code{DFmode}, @code{XFmode} or
907 @code{TFmode} floating point numbers are stored in memory with the word
908 containing the sign bit at the lowest address; otherwise define it to
909 have the value 0. This macro need not be a constant.
911 You need not define this macro if the ordering is the same as for
915 @defmac BITS_PER_UNIT
916 Define this macro to be the number of bits in an addressable storage
917 unit (byte). If you do not define this macro the default is 8.
920 @defmac BITS_PER_WORD
921 Number of bits in a word. If you do not define this macro, the default
922 is @code{BITS_PER_UNIT * UNITS_PER_WORD}.
925 @defmac MAX_BITS_PER_WORD
926 Maximum number of bits in a word. If this is undefined, the default is
927 @code{BITS_PER_WORD}. Otherwise, it is the constant value that is the
928 largest value that @code{BITS_PER_WORD} can have at run-time.
931 @defmac UNITS_PER_WORD
932 Number of storage units in a word; normally the size of a general-purpose
933 register, a power of two from 1 or 8.
936 @defmac MIN_UNITS_PER_WORD
937 Minimum number of units in a word. If this is undefined, the default is
938 @code{UNITS_PER_WORD}. Otherwise, it is the constant value that is the
939 smallest value that @code{UNITS_PER_WORD} can have at run-time.
942 @defmac UNITS_PER_SIMD_WORD
943 Number of units in the vectors that the vectorizer can produce.
944 The default is equal to @code{UNITS_PER_WORD}, because the vectorizer
945 can do some transformations even in absence of specialized @acronym{SIMD}
950 Width of a pointer, in bits. You must specify a value no wider than the
951 width of @code{Pmode}. If it is not equal to the width of @code{Pmode},
952 you must define @code{POINTERS_EXTEND_UNSIGNED}. If you do not specify
953 a value the default is @code{BITS_PER_WORD}.
956 @defmac POINTERS_EXTEND_UNSIGNED
957 A C expression whose value is greater than zero if pointers that need to be
958 extended from being @code{POINTER_SIZE} bits wide to @code{Pmode} are to
959 be zero-extended and zero if they are to be sign-extended. If the value
960 is less then zero then there must be an "ptr_extend" instruction that
961 extends a pointer from @code{POINTER_SIZE} to @code{Pmode}.
963 You need not define this macro if the @code{POINTER_SIZE} is equal
964 to the width of @code{Pmode}.
967 @defmac PROMOTE_MODE (@var{m}, @var{unsignedp}, @var{type})
968 A macro to update @var{m} and @var{unsignedp} when an object whose type
969 is @var{type} and which has the specified mode and signedness is to be
970 stored in a register. This macro is only called when @var{type} is a
973 On most RISC machines, which only have operations that operate on a full
974 register, define this macro to set @var{m} to @code{word_mode} if
975 @var{m} is an integer mode narrower than @code{BITS_PER_WORD}. In most
976 cases, only integer modes should be widened because wider-precision
977 floating-point operations are usually more expensive than their narrower
980 For most machines, the macro definition does not change @var{unsignedp}.
981 However, some machines, have instructions that preferentially handle
982 either signed or unsigned quantities of certain modes. For example, on
983 the DEC Alpha, 32-bit loads from memory and 32-bit add instructions
984 sign-extend the result to 64 bits. On such machines, set
985 @var{unsignedp} according to which kind of extension is more efficient.
987 Do not define this macro if it would never modify @var{m}.
990 @defmac PROMOTE_FUNCTION_MODE
991 Like @code{PROMOTE_MODE}, but is applied to outgoing function arguments or
992 function return values, as specified by @code{TARGET_PROMOTE_FUNCTION_ARGS}
993 and @code{TARGET_PROMOTE_FUNCTION_RETURN}, respectively.
995 The default is @code{PROMOTE_MODE}.
998 @deftypefn {Target Hook} bool TARGET_PROMOTE_FUNCTION_ARGS (tree @var{fntype})
999 This target hook should return @code{true} if the promotion described by
1000 @code{PROMOTE_FUNCTION_MODE} should be done for outgoing function
1004 @deftypefn {Target Hook} bool TARGET_PROMOTE_FUNCTION_RETURN (tree @var{fntype})
1005 This target hook should return @code{true} if the promotion described by
1006 @code{PROMOTE_FUNCTION_MODE} should be done for the return value of
1009 If this target hook returns @code{true}, @code{FUNCTION_VALUE} must
1010 perform the same promotions done by @code{PROMOTE_FUNCTION_MODE}.
1013 @defmac PARM_BOUNDARY
1014 Normal alignment required for function parameters on the stack, in
1015 bits. All stack parameters receive at least this much alignment
1016 regardless of data type. On most machines, this is the same as the
1020 @defmac STACK_BOUNDARY
1021 Define this macro to the minimum alignment enforced by hardware for the
1022 stack pointer on this machine. The definition is a C expression for the
1023 desired alignment (measured in bits). This value is used as a default
1024 if @code{PREFERRED_STACK_BOUNDARY} is not defined. On most machines,
1025 this should be the same as @code{PARM_BOUNDARY}.
1028 @defmac PREFERRED_STACK_BOUNDARY
1029 Define this macro if you wish to preserve a certain alignment for the
1030 stack pointer, greater than what the hardware enforces. The definition
1031 is a C expression for the desired alignment (measured in bits). This
1032 macro must evaluate to a value equal to or larger than
1033 @code{STACK_BOUNDARY}.
1036 @defmac FUNCTION_BOUNDARY
1037 Alignment required for a function entry point, in bits.
1040 @defmac BIGGEST_ALIGNMENT
1041 Biggest alignment that any data type can require on this machine, in bits.
1044 @defmac MINIMUM_ATOMIC_ALIGNMENT
1045 If defined, the smallest alignment, in bits, that can be given to an
1046 object that can be referenced in one operation, without disturbing any
1047 nearby object. Normally, this is @code{BITS_PER_UNIT}, but may be larger
1048 on machines that don't have byte or half-word store operations.
1051 @defmac BIGGEST_FIELD_ALIGNMENT
1052 Biggest alignment that any structure or union field can require on this
1053 machine, in bits. If defined, this overrides @code{BIGGEST_ALIGNMENT} for
1054 structure and union fields only, unless the field alignment has been set
1055 by the @code{__attribute__ ((aligned (@var{n})))} construct.
1058 @defmac ADJUST_FIELD_ALIGN (@var{field}, @var{computed})
1059 An expression for the alignment of a structure field @var{field} if the
1060 alignment computed in the usual way (including applying of
1061 @code{BIGGEST_ALIGNMENT} and @code{BIGGEST_FIELD_ALIGNMENT} to the
1062 alignment) is @var{computed}. It overrides alignment only if the
1063 field alignment has not been set by the
1064 @code{__attribute__ ((aligned (@var{n})))} construct.
1067 @defmac MAX_OFILE_ALIGNMENT
1068 Biggest alignment supported by the object file format of this machine.
1069 Use this macro to limit the alignment which can be specified using the
1070 @code{__attribute__ ((aligned (@var{n})))} construct. If not defined,
1071 the default value is @code{BIGGEST_ALIGNMENT}.
1074 @defmac DATA_ALIGNMENT (@var{type}, @var{basic-align})
1075 If defined, a C expression to compute the alignment for a variable in
1076 the static store. @var{type} is the data type, and @var{basic-align} is
1077 the alignment that the object would ordinarily have. The value of this
1078 macro is used instead of that alignment to align the object.
1080 If this macro is not defined, then @var{basic-align} is used.
1083 One use of this macro is to increase alignment of medium-size data to
1084 make it all fit in fewer cache lines. Another is to cause character
1085 arrays to be word-aligned so that @code{strcpy} calls that copy
1086 constants to character arrays can be done inline.
1089 @defmac CONSTANT_ALIGNMENT (@var{constant}, @var{basic-align})
1090 If defined, a C expression to compute the alignment given to a constant
1091 that is being placed in memory. @var{constant} is the constant and
1092 @var{basic-align} is the alignment that the object would ordinarily
1093 have. The value of this macro is used instead of that alignment to
1096 If this macro is not defined, then @var{basic-align} is used.
1098 The typical use of this macro is to increase alignment for string
1099 constants to be word aligned so that @code{strcpy} calls that copy
1100 constants can be done inline.
1103 @defmac LOCAL_ALIGNMENT (@var{type}, @var{basic-align})
1104 If defined, a C expression to compute the alignment for a variable in
1105 the local store. @var{type} is the data type, and @var{basic-align} is
1106 the alignment that the object would ordinarily have. The value of this
1107 macro is used instead of that alignment to align the object.
1109 If this macro is not defined, then @var{basic-align} is used.
1111 One use of this macro is to increase alignment of medium-size data to
1112 make it all fit in fewer cache lines.
1115 @defmac EMPTY_FIELD_BOUNDARY
1116 Alignment in bits to be given to a structure bit-field that follows an
1117 empty field such as @code{int : 0;}.
1119 If @code{PCC_BITFIELD_TYPE_MATTERS} is true, it overrides this macro.
1122 @defmac STRUCTURE_SIZE_BOUNDARY
1123 Number of bits which any structure or union's size must be a multiple of.
1124 Each structure or union's size is rounded up to a multiple of this.
1126 If you do not define this macro, the default is the same as
1127 @code{BITS_PER_UNIT}.
1130 @defmac STRICT_ALIGNMENT
1131 Define this macro to be the value 1 if instructions will fail to work
1132 if given data not on the nominal alignment. If instructions will merely
1133 go slower in that case, define this macro as 0.
1136 @defmac PCC_BITFIELD_TYPE_MATTERS
1137 Define this if you wish to imitate the way many other C compilers handle
1138 alignment of bit-fields and the structures that contain them.
1140 The behavior is that the type written for a named bit-field (@code{int},
1141 @code{short}, or other integer type) imposes an alignment for the entire
1142 structure, as if the structure really did contain an ordinary field of
1143 that type. In addition, the bit-field is placed within the structure so
1144 that it would fit within such a field, not crossing a boundary for it.
1146 Thus, on most machines, a named bit-field whose type is written as
1147 @code{int} would not cross a four-byte boundary, and would force
1148 four-byte alignment for the whole structure. (The alignment used may
1149 not be four bytes; it is controlled by the other alignment parameters.)
1151 An unnamed bit-field will not affect the alignment of the containing
1154 If the macro is defined, its definition should be a C expression;
1155 a nonzero value for the expression enables this behavior.
1157 Note that if this macro is not defined, or its value is zero, some
1158 bit-fields may cross more than one alignment boundary. The compiler can
1159 support such references if there are @samp{insv}, @samp{extv}, and
1160 @samp{extzv} insns that can directly reference memory.
1162 The other known way of making bit-fields work is to define
1163 @code{STRUCTURE_SIZE_BOUNDARY} as large as @code{BIGGEST_ALIGNMENT}.
1164 Then every structure can be accessed with fullwords.
1166 Unless the machine has bit-field instructions or you define
1167 @code{STRUCTURE_SIZE_BOUNDARY} that way, you must define
1168 @code{PCC_BITFIELD_TYPE_MATTERS} to have a nonzero value.
1170 If your aim is to make GCC use the same conventions for laying out
1171 bit-fields as are used by another compiler, here is how to investigate
1172 what the other compiler does. Compile and run this program:
1191 printf ("Size of foo1 is %d\n",
1192 sizeof (struct foo1));
1193 printf ("Size of foo2 is %d\n",
1194 sizeof (struct foo2));
1199 If this prints 2 and 5, then the compiler's behavior is what you would
1200 get from @code{PCC_BITFIELD_TYPE_MATTERS}.
1203 @defmac BITFIELD_NBYTES_LIMITED
1204 Like @code{PCC_BITFIELD_TYPE_MATTERS} except that its effect is limited
1205 to aligning a bit-field within the structure.
1208 @deftypefn {Target Hook} bool TARGET_ALIGN_ANON_BITFIELDS (void)
1209 When @code{PCC_BITFIELD_TYPE_MATTERS} is true this hook will determine
1210 whether unnamed bitfields affect the alignment of the containing
1211 structure. The hook should return true if the structure should inherit
1212 the alignment requirements of an unnamed bitfield's type.
1215 @defmac MEMBER_TYPE_FORCES_BLK (@var{field}, @var{mode})
1216 Return 1 if a structure or array containing @var{field} should be accessed using
1219 If @var{field} is the only field in the structure, @var{mode} is its
1220 mode, otherwise @var{mode} is VOIDmode. @var{mode} is provided in the
1221 case where structures of one field would require the structure's mode to
1222 retain the field's mode.
1224 Normally, this is not needed. See the file @file{c4x.h} for an example
1225 of how to use this macro to prevent a structure having a floating point
1226 field from being accessed in an integer mode.
1229 @defmac ROUND_TYPE_ALIGN (@var{type}, @var{computed}, @var{specified})
1230 Define this macro as an expression for the alignment of a type (given
1231 by @var{type} as a tree node) if the alignment computed in the usual
1232 way is @var{computed} and the alignment explicitly specified was
1235 The default is to use @var{specified} if it is larger; otherwise, use
1236 the smaller of @var{computed} and @code{BIGGEST_ALIGNMENT}
1239 @defmac MAX_FIXED_MODE_SIZE
1240 An integer expression for the size in bits of the largest integer
1241 machine mode that should actually be used. All integer machine modes of
1242 this size or smaller can be used for structures and unions with the
1243 appropriate sizes. If this macro is undefined, @code{GET_MODE_BITSIZE
1244 (DImode)} is assumed.
1247 @defmac STACK_SAVEAREA_MODE (@var{save_level})
1248 If defined, an expression of type @code{enum machine_mode} that
1249 specifies the mode of the save area operand of a
1250 @code{save_stack_@var{level}} named pattern (@pxref{Standard Names}).
1251 @var{save_level} is one of @code{SAVE_BLOCK}, @code{SAVE_FUNCTION}, or
1252 @code{SAVE_NONLOCAL} and selects which of the three named patterns is
1253 having its mode specified.
1255 You need not define this macro if it always returns @code{Pmode}. You
1256 would most commonly define this macro if the
1257 @code{save_stack_@var{level}} patterns need to support both a 32- and a
1261 @defmac STACK_SIZE_MODE
1262 If defined, an expression of type @code{enum machine_mode} that
1263 specifies the mode of the size increment operand of an
1264 @code{allocate_stack} named pattern (@pxref{Standard Names}).
1266 You need not define this macro if it always returns @code{word_mode}.
1267 You would most commonly define this macro if the @code{allocate_stack}
1268 pattern needs to support both a 32- and a 64-bit mode.
1271 @defmac TARGET_FLOAT_FORMAT
1272 A code distinguishing the floating point format of the target machine.
1273 There are four defined values:
1276 @item IEEE_FLOAT_FORMAT
1277 This code indicates IEEE floating point. It is the default; there is no
1278 need to define @code{TARGET_FLOAT_FORMAT} when the format is IEEE@.
1280 @item VAX_FLOAT_FORMAT
1281 This code indicates the ``F float'' (for @code{float}) and ``D float''
1282 or ``G float'' formats (for @code{double}) used on the VAX and PDP-11@.
1284 @item IBM_FLOAT_FORMAT
1285 This code indicates the format used on the IBM System/370.
1287 @item C4X_FLOAT_FORMAT
1288 This code indicates the format used on the TMS320C3x/C4x.
1291 If your target uses a floating point format other than these, you must
1292 define a new @var{name}_FLOAT_FORMAT code for it, and add support for
1293 it to @file{real.c}.
1295 The ordering of the component words of floating point values stored in
1296 memory is controlled by @code{FLOAT_WORDS_BIG_ENDIAN}.
1299 @defmac MODE_HAS_NANS (@var{mode})
1300 When defined, this macro should be true if @var{mode} has a NaN
1301 representation. The compiler assumes that NaNs are not equal to
1302 anything (including themselves) and that addition, subtraction,
1303 multiplication and division all return NaNs when one operand is
1306 By default, this macro is true if @var{mode} is a floating-point
1307 mode and the target floating-point format is IEEE@.
1310 @defmac MODE_HAS_INFINITIES (@var{mode})
1311 This macro should be true if @var{mode} can represent infinity. At
1312 present, the compiler uses this macro to decide whether @samp{x - x}
1313 is always defined. By default, the macro is true when @var{mode}
1314 is a floating-point mode and the target format is IEEE@.
1317 @defmac MODE_HAS_SIGNED_ZEROS (@var{mode})
1318 True if @var{mode} distinguishes between positive and negative zero.
1319 The rules are expected to follow the IEEE standard:
1323 @samp{x + x} has the same sign as @samp{x}.
1326 If the sum of two values with opposite sign is zero, the result is
1327 positive for all rounding modes expect towards @minus{}infinity, for
1328 which it is negative.
1331 The sign of a product or quotient is negative when exactly one
1332 of the operands is negative.
1335 The default definition is true if @var{mode} is a floating-point
1336 mode and the target format is IEEE@.
1339 @defmac MODE_HAS_SIGN_DEPENDENT_ROUNDING (@var{mode})
1340 If defined, this macro should be true for @var{mode} if it has at
1341 least one rounding mode in which @samp{x} and @samp{-x} can be
1342 rounded to numbers of different magnitude. Two such modes are
1343 towards @minus{}infinity and towards +infinity.
1345 The default definition of this macro is true if @var{mode} is
1346 a floating-point mode and the target format is IEEE@.
1349 @defmac ROUND_TOWARDS_ZERO
1350 If defined, this macro should be true if the prevailing rounding
1351 mode is towards zero. A true value has the following effects:
1355 @code{MODE_HAS_SIGN_DEPENDENT_ROUNDING} will be false for all modes.
1358 @file{libgcc.a}'s floating-point emulator will round towards zero
1359 rather than towards nearest.
1362 The compiler's floating-point emulator will round towards zero after
1363 doing arithmetic, and when converting from the internal float format to
1367 The macro does not affect the parsing of string literals. When the
1368 primary rounding mode is towards zero, library functions like
1369 @code{strtod} might still round towards nearest, and the compiler's
1370 parser should behave like the target's @code{strtod} where possible.
1372 Not defining this macro is equivalent to returning zero.
1375 @defmac LARGEST_EXPONENT_IS_NORMAL (@var{size})
1376 This macro should return true if floats with @var{size}
1377 bits do not have a NaN or infinity representation, but use the largest
1378 exponent for normal numbers instead.
1380 Defining this macro to true for @var{size} causes @code{MODE_HAS_NANS}
1381 and @code{MODE_HAS_INFINITIES} to be false for @var{size}-bit modes.
1382 It also affects the way @file{libgcc.a} and @file{real.c} emulate
1383 floating-point arithmetic.
1385 The default definition of this macro returns false for all sizes.
1388 @deftypefn {Target Hook} bool TARGET_VECTOR_OPAQUE_P (tree @var{type})
1389 This target hook should return @code{true} a vector is opaque. That
1390 is, if no cast is needed when copying a vector value of type
1391 @var{type} into another vector lvalue of the same size. Vector opaque
1392 types cannot be initialized. The default is that there are no such
1396 @deftypefn {Target Hook} bool TARGET_MS_BITFIELD_LAYOUT_P (tree @var{record_type})
1397 This target hook returns @code{true} if bit-fields in the given
1398 @var{record_type} are to be laid out following the rules of Microsoft
1399 Visual C/C++, namely: (i) a bit-field won't share the same storage
1400 unit with the previous bit-field if their underlying types have
1401 different sizes, and the bit-field will be aligned to the highest
1402 alignment of the underlying types of itself and of the previous
1403 bit-field; (ii) a zero-sized bit-field will affect the alignment of
1404 the whole enclosing structure, even if it is unnamed; except that
1405 (iii) a zero-sized bit-field will be disregarded unless it follows
1406 another bit-field of nonzero size. If this hook returns @code{true},
1407 other macros that control bit-field layout are ignored.
1409 When a bit-field is inserted into a packed record, the whole size
1410 of the underlying type is used by one or more same-size adjacent
1411 bit-fields (that is, if its long:3, 32 bits is used in the record,
1412 and any additional adjacent long bit-fields are packed into the same
1413 chunk of 32 bits. However, if the size changes, a new field of that
1414 size is allocated). In an unpacked record, this is the same as using
1415 alignment, but not equivalent when packing.
1417 If both MS bit-fields and @samp{__attribute__((packed))} are used,
1418 the latter will take precedence. If @samp{__attribute__((packed))} is
1419 used on a single field when MS bit-fields are in use, it will take
1420 precedence for that field, but the alignment of the rest of the structure
1421 may affect its placement.
1424 @deftypefn {Target Hook} {const char *} TARGET_MANGLE_FUNDAMENTAL_TYPE (tree @var{type})
1425 If your target defines any fundamental types, define this hook to
1426 return the appropriate encoding for these types as part of a C++
1427 mangled name. The @var{type} argument is the tree structure
1428 representing the type to be mangled. The hook may be applied to trees
1429 which are not target-specific fundamental types; it should return
1430 @code{NULL} for all such types, as well as arguments it does not
1431 recognize. If the return value is not @code{NULL}, it must point to
1432 a statically-allocated string constant.
1434 Target-specific fundamental types might be new fundamental types or
1435 qualified versions of ordinary fundamental types. Encode new
1436 fundamental types as @samp{@w{u @var{n} @var{name}}}, where @var{name}
1437 is the name used for the type in source code, and @var{n} is the
1438 length of @var{name} in decimal. Encode qualified versions of
1439 ordinary types as @samp{@w{U @var{n} @var{name} @var{code}}}, where
1440 @var{name} is the name used for the type qualifier in source code,
1441 @var{n} is the length of @var{name} as above, and @var{code} is the
1442 code used to represent the unqualified version of this type. (See
1443 @code{write_builtin_type} in @file{cp/mangle.c} for the list of
1444 codes.) In both cases the spaces are for clarity; do not include any
1445 spaces in your string.
1447 The default version of this hook always returns @code{NULL}, which is
1448 appropriate for a target that does not define any new fundamental
1453 @section Layout of Source Language Data Types
1455 These macros define the sizes and other characteristics of the standard
1456 basic data types used in programs being compiled. Unlike the macros in
1457 the previous section, these apply to specific features of C and related
1458 languages, rather than to fundamental aspects of storage layout.
1460 @defmac INT_TYPE_SIZE
1461 A C expression for the size in bits of the type @code{int} on the
1462 target machine. If you don't define this, the default is one word.
1465 @defmac SHORT_TYPE_SIZE
1466 A C expression for the size in bits of the type @code{short} on the
1467 target machine. If you don't define this, the default is half a word.
1468 (If this would be less than one storage unit, it is rounded up to one
1472 @defmac LONG_TYPE_SIZE
1473 A C expression for the size in bits of the type @code{long} on the
1474 target machine. If you don't define this, the default is one word.
1477 @defmac ADA_LONG_TYPE_SIZE
1478 On some machines, the size used for the Ada equivalent of the type
1479 @code{long} by a native Ada compiler differs from that used by C@. In
1480 that situation, define this macro to be a C expression to be used for
1481 the size of that type. If you don't define this, the default is the
1482 value of @code{LONG_TYPE_SIZE}.
1485 @defmac LONG_LONG_TYPE_SIZE
1486 A C expression for the size in bits of the type @code{long long} on the
1487 target machine. If you don't define this, the default is two
1488 words. If you want to support GNU Ada on your machine, the value of this
1489 macro must be at least 64.
1492 @defmac CHAR_TYPE_SIZE
1493 A C expression for the size in bits of the type @code{char} on the
1494 target machine. If you don't define this, the default is
1495 @code{BITS_PER_UNIT}.
1498 @defmac BOOL_TYPE_SIZE
1499 A C expression for the size in bits of the C++ type @code{bool} and
1500 C99 type @code{_Bool} on the target machine. If you don't define
1501 this, and you probably shouldn't, the default is @code{CHAR_TYPE_SIZE}.
1504 @defmac FLOAT_TYPE_SIZE
1505 A C expression for the size in bits of the type @code{float} on the
1506 target machine. If you don't define this, the default is one word.
1509 @defmac DOUBLE_TYPE_SIZE
1510 A C expression for the size in bits of the type @code{double} on the
1511 target machine. If you don't define this, the default is two
1515 @defmac LONG_DOUBLE_TYPE_SIZE
1516 A C expression for the size in bits of the type @code{long double} on
1517 the target machine. If you don't define this, the default is two
1521 @defmac LIBGCC2_LONG_DOUBLE_TYPE_SIZE
1522 Define this macro if @code{LONG_DOUBLE_TYPE_SIZE} is not constant or
1523 if you want routines in @file{libgcc2.a} for a size other than
1524 @code{LONG_DOUBLE_TYPE_SIZE}. If you don't define this, the
1525 default is @code{LONG_DOUBLE_TYPE_SIZE}.
1528 @defmac LIBGCC2_HAS_DF_MODE
1529 Define this macro if neither @code{LIBGCC2_DOUBLE_TYPE_SIZE} nor
1530 @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is
1531 @code{DFmode} but you want @code{DFmode} routines in @file{libgcc2.a}
1532 anyway. If you don't define this and either @code{LIBGCC2_DOUBLE_TYPE_SIZE}
1533 or @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is 64 then the default is 1,
1537 @defmac LIBGCC2_HAS_XF_MODE
1538 Define this macro if @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is not
1539 @code{XFmode} but you want @code{XFmode} routines in @file{libgcc2.a}
1540 anyway. If you don't define this and @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE}
1541 is 80 then the default is 1, otherwise it is 0.
1544 @defmac LIBGCC2_HAS_TF_MODE
1545 Define this macro if @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is not
1546 @code{TFmode} but you want @code{TFmode} routines in @file{libgcc2.a}
1547 anyway. If you don't define this and @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE}
1548 is 128 then the default is 1, otherwise it is 0.
1551 @defmac TARGET_FLT_EVAL_METHOD
1552 A C expression for the value for @code{FLT_EVAL_METHOD} in @file{float.h},
1553 assuming, if applicable, that the floating-point control word is in its
1554 default state. If you do not define this macro the value of
1555 @code{FLT_EVAL_METHOD} will be zero.
1558 @defmac WIDEST_HARDWARE_FP_SIZE
1559 A C expression for the size in bits of the widest floating-point format
1560 supported by the hardware. If you define this macro, you must specify a
1561 value less than or equal to the value of @code{LONG_DOUBLE_TYPE_SIZE}.
1562 If you do not define this macro, the value of @code{LONG_DOUBLE_TYPE_SIZE}
1566 @defmac DEFAULT_SIGNED_CHAR
1567 An expression whose value is 1 or 0, according to whether the type
1568 @code{char} should be signed or unsigned by default. The user can
1569 always override this default with the options @option{-fsigned-char}
1570 and @option{-funsigned-char}.
1573 @deftypefn {Target Hook} bool TARGET_DEFAULT_SHORT_ENUMS (void)
1574 This target hook should return true if the compiler should give an
1575 @code{enum} type only as many bytes as it takes to represent the range
1576 of possible values of that type. It should return false if all
1577 @code{enum} types should be allocated like @code{int}.
1579 The default is to return false.
1583 A C expression for a string describing the name of the data type to use
1584 for size values. The typedef name @code{size_t} is defined using the
1585 contents of the string.
1587 The string can contain more than one keyword. If so, separate them with
1588 spaces, and write first any length keyword, then @code{unsigned} if
1589 appropriate, and finally @code{int}. The string must exactly match one
1590 of the data type names defined in the function
1591 @code{init_decl_processing} in the file @file{c-decl.c}. You may not
1592 omit @code{int} or change the order---that would cause the compiler to
1595 If you don't define this macro, the default is @code{"long unsigned
1599 @defmac PTRDIFF_TYPE
1600 A C expression for a string describing the name of the data type to use
1601 for the result of subtracting two pointers. The typedef name
1602 @code{ptrdiff_t} is defined using the contents of the string. See
1603 @code{SIZE_TYPE} above for more information.
1605 If you don't define this macro, the default is @code{"long int"}.
1609 A C expression for a string describing the name of the data type to use
1610 for wide characters. The typedef name @code{wchar_t} is defined using
1611 the contents of the string. See @code{SIZE_TYPE} above for more
1614 If you don't define this macro, the default is @code{"int"}.
1617 @defmac WCHAR_TYPE_SIZE
1618 A C expression for the size in bits of the data type for wide
1619 characters. This is used in @code{cpp}, which cannot make use of
1624 A C expression for a string describing the name of the data type to
1625 use for wide characters passed to @code{printf} and returned from
1626 @code{getwc}. The typedef name @code{wint_t} is defined using the
1627 contents of the string. See @code{SIZE_TYPE} above for more
1630 If you don't define this macro, the default is @code{"unsigned int"}.
1634 A C expression for a string describing the name of the data type that
1635 can represent any value of any standard or extended signed integer type.
1636 The typedef name @code{intmax_t} is defined using the contents of the
1637 string. See @code{SIZE_TYPE} above for more information.
1639 If you don't define this macro, the default is the first of
1640 @code{"int"}, @code{"long int"}, or @code{"long long int"} that has as
1641 much precision as @code{long long int}.
1644 @defmac UINTMAX_TYPE
1645 A C expression for a string describing the name of the data type that
1646 can represent any value of any standard or extended unsigned integer
1647 type. The typedef name @code{uintmax_t} is defined using the contents
1648 of the string. See @code{SIZE_TYPE} above for more information.
1650 If you don't define this macro, the default is the first of
1651 @code{"unsigned int"}, @code{"long unsigned int"}, or @code{"long long
1652 unsigned int"} that has as much precision as @code{long long unsigned
1656 @defmac TARGET_PTRMEMFUNC_VBIT_LOCATION
1657 The C++ compiler represents a pointer-to-member-function with a struct
1664 ptrdiff_t vtable_index;
1671 The C++ compiler must use one bit to indicate whether the function that
1672 will be called through a pointer-to-member-function is virtual.
1673 Normally, we assume that the low-order bit of a function pointer must
1674 always be zero. Then, by ensuring that the vtable_index is odd, we can
1675 distinguish which variant of the union is in use. But, on some
1676 platforms function pointers can be odd, and so this doesn't work. In
1677 that case, we use the low-order bit of the @code{delta} field, and shift
1678 the remainder of the @code{delta} field to the left.
1680 GCC will automatically make the right selection about where to store
1681 this bit using the @code{FUNCTION_BOUNDARY} setting for your platform.
1682 However, some platforms such as ARM/Thumb have @code{FUNCTION_BOUNDARY}
1683 set such that functions always start at even addresses, but the lowest
1684 bit of pointers to functions indicate whether the function at that
1685 address is in ARM or Thumb mode. If this is the case of your
1686 architecture, you should define this macro to
1687 @code{ptrmemfunc_vbit_in_delta}.
1689 In general, you should not have to define this macro. On architectures
1690 in which function addresses are always even, according to
1691 @code{FUNCTION_BOUNDARY}, GCC will automatically define this macro to
1692 @code{ptrmemfunc_vbit_in_pfn}.
1695 @defmac TARGET_VTABLE_USES_DESCRIPTORS
1696 Normally, the C++ compiler uses function pointers in vtables. This
1697 macro allows the target to change to use ``function descriptors''
1698 instead. Function descriptors are found on targets for whom a
1699 function pointer is actually a small data structure. Normally the
1700 data structure consists of the actual code address plus a data
1701 pointer to which the function's data is relative.
1703 If vtables are used, the value of this macro should be the number
1704 of words that the function descriptor occupies.
1707 @defmac TARGET_VTABLE_ENTRY_ALIGN
1708 By default, the vtable entries are void pointers, the so the alignment
1709 is the same as pointer alignment. The value of this macro specifies
1710 the alignment of the vtable entry in bits. It should be defined only
1711 when special alignment is necessary. */
1714 @defmac TARGET_VTABLE_DATA_ENTRY_DISTANCE
1715 There are a few non-descriptor entries in the vtable at offsets below
1716 zero. If these entries must be padded (say, to preserve the alignment
1717 specified by @code{TARGET_VTABLE_ENTRY_ALIGN}), set this to the number
1718 of words in each data entry.
1722 @section Register Usage
1723 @cindex register usage
1725 This section explains how to describe what registers the target machine
1726 has, and how (in general) they can be used.
1728 The description of which registers a specific instruction can use is
1729 done with register classes; see @ref{Register Classes}. For information
1730 on using registers to access a stack frame, see @ref{Frame Registers}.
1731 For passing values in registers, see @ref{Register Arguments}.
1732 For returning values in registers, see @ref{Scalar Return}.
1735 * Register Basics:: Number and kinds of registers.
1736 * Allocation Order:: Order in which registers are allocated.
1737 * Values in Registers:: What kinds of values each reg can hold.
1738 * Leaf Functions:: Renumbering registers for leaf functions.
1739 * Stack Registers:: Handling a register stack such as 80387.
1742 @node Register Basics
1743 @subsection Basic Characteristics of Registers
1745 @c prevent bad page break with this line
1746 Registers have various characteristics.
1748 @defmac FIRST_PSEUDO_REGISTER
1749 Number of hardware registers known to the compiler. They receive
1750 numbers 0 through @code{FIRST_PSEUDO_REGISTER-1}; thus, the first
1751 pseudo register's number really is assigned the number
1752 @code{FIRST_PSEUDO_REGISTER}.
1755 @defmac FIXED_REGISTERS
1756 @cindex fixed register
1757 An initializer that says which registers are used for fixed purposes
1758 all throughout the compiled code and are therefore not available for
1759 general allocation. These would include the stack pointer, the frame
1760 pointer (except on machines where that can be used as a general
1761 register when no frame pointer is needed), the program counter on
1762 machines where that is considered one of the addressable registers,
1763 and any other numbered register with a standard use.
1765 This information is expressed as a sequence of numbers, separated by
1766 commas and surrounded by braces. The @var{n}th number is 1 if
1767 register @var{n} is fixed, 0 otherwise.
1769 The table initialized from this macro, and the table initialized by
1770 the following one, may be overridden at run time either automatically,
1771 by the actions of the macro @code{CONDITIONAL_REGISTER_USAGE}, or by
1772 the user with the command options @option{-ffixed-@var{reg}},
1773 @option{-fcall-used-@var{reg}} and @option{-fcall-saved-@var{reg}}.
1776 @defmac CALL_USED_REGISTERS
1777 @cindex call-used register
1778 @cindex call-clobbered register
1779 @cindex call-saved register
1780 Like @code{FIXED_REGISTERS} but has 1 for each register that is
1781 clobbered (in general) by function calls as well as for fixed
1782 registers. This macro therefore identifies the registers that are not
1783 available for general allocation of values that must live across
1786 If a register has 0 in @code{CALL_USED_REGISTERS}, the compiler
1787 automatically saves it on function entry and restores it on function
1788 exit, if the register is used within the function.
1791 @defmac CALL_REALLY_USED_REGISTERS
1792 @cindex call-used register
1793 @cindex call-clobbered register
1794 @cindex call-saved register
1795 Like @code{CALL_USED_REGISTERS} except this macro doesn't require
1796 that the entire set of @code{FIXED_REGISTERS} be included.
1797 (@code{CALL_USED_REGISTERS} must be a superset of @code{FIXED_REGISTERS}).
1798 This macro is optional. If not specified, it defaults to the value
1799 of @code{CALL_USED_REGISTERS}.
1802 @defmac HARD_REGNO_CALL_PART_CLOBBERED (@var{regno}, @var{mode})
1803 @cindex call-used register
1804 @cindex call-clobbered register
1805 @cindex call-saved register
1806 A C expression that is nonzero if it is not permissible to store a
1807 value of mode @var{mode} in hard register number @var{regno} across a
1808 call without some part of it being clobbered. For most machines this
1809 macro need not be defined. It is only required for machines that do not
1810 preserve the entire contents of a register across a call.
1814 @findex call_used_regs
1817 @findex reg_class_contents
1818 @defmac CONDITIONAL_REGISTER_USAGE
1819 Zero or more C statements that may conditionally modify five variables
1820 @code{fixed_regs}, @code{call_used_regs}, @code{global_regs},
1821 @code{reg_names}, and @code{reg_class_contents}, to take into account
1822 any dependence of these register sets on target flags. The first three
1823 of these are of type @code{char []} (interpreted as Boolean vectors).
1824 @code{global_regs} is a @code{const char *[]}, and
1825 @code{reg_class_contents} is a @code{HARD_REG_SET}. Before the macro is
1826 called, @code{fixed_regs}, @code{call_used_regs},
1827 @code{reg_class_contents}, and @code{reg_names} have been initialized
1828 from @code{FIXED_REGISTERS}, @code{CALL_USED_REGISTERS},
1829 @code{REG_CLASS_CONTENTS}, and @code{REGISTER_NAMES}, respectively.
1830 @code{global_regs} has been cleared, and any @option{-ffixed-@var{reg}},
1831 @option{-fcall-used-@var{reg}} and @option{-fcall-saved-@var{reg}}
1832 command options have been applied.
1834 You need not define this macro if it has no work to do.
1836 @cindex disabling certain registers
1837 @cindex controlling register usage
1838 If the usage of an entire class of registers depends on the target
1839 flags, you may indicate this to GCC by using this macro to modify
1840 @code{fixed_regs} and @code{call_used_regs} to 1 for each of the
1841 registers in the classes which should not be used by GCC@. Also define
1842 the macro @code{REG_CLASS_FROM_LETTER} / @code{REG_CLASS_FROM_CONSTRAINT}
1843 to return @code{NO_REGS} if it
1844 is called with a letter for a class that shouldn't be used.
1846 (However, if this class is not included in @code{GENERAL_REGS} and all
1847 of the insn patterns whose constraints permit this class are
1848 controlled by target switches, then GCC will automatically avoid using
1849 these registers when the target switches are opposed to them.)
1852 @defmac INCOMING_REGNO (@var{out})
1853 Define this macro if the target machine has register windows. This C
1854 expression returns the register number as seen by the called function
1855 corresponding to the register number @var{out} as seen by the calling
1856 function. Return @var{out} if register number @var{out} is not an
1860 @defmac OUTGOING_REGNO (@var{in})
1861 Define this macro if the target machine has register windows. This C
1862 expression returns the register number as seen by the calling function
1863 corresponding to the register number @var{in} as seen by the called
1864 function. Return @var{in} if register number @var{in} is not an inbound
1868 @defmac LOCAL_REGNO (@var{regno})
1869 Define this macro if the target machine has register windows. This C
1870 expression returns true if the register is call-saved but is in the
1871 register window. Unlike most call-saved registers, such registers
1872 need not be explicitly restored on function exit or during non-local
1877 If the program counter has a register number, define this as that
1878 register number. Otherwise, do not define it.
1881 @node Allocation Order
1882 @subsection Order of Allocation of Registers
1883 @cindex order of register allocation
1884 @cindex register allocation order
1886 @c prevent bad page break with this line
1887 Registers are allocated in order.
1889 @defmac REG_ALLOC_ORDER
1890 If defined, an initializer for a vector of integers, containing the
1891 numbers of hard registers in the order in which GCC should prefer
1892 to use them (from most preferred to least).
1894 If this macro is not defined, registers are used lowest numbered first
1895 (all else being equal).
1897 One use of this macro is on machines where the highest numbered
1898 registers must always be saved and the save-multiple-registers
1899 instruction supports only sequences of consecutive registers. On such
1900 machines, define @code{REG_ALLOC_ORDER} to be an initializer that lists
1901 the highest numbered allocable register first.
1904 @defmac ORDER_REGS_FOR_LOCAL_ALLOC
1905 A C statement (sans semicolon) to choose the order in which to allocate
1906 hard registers for pseudo-registers local to a basic block.
1908 Store the desired register order in the array @code{reg_alloc_order}.
1909 Element 0 should be the register to allocate first; element 1, the next
1910 register; and so on.
1912 The macro body should not assume anything about the contents of
1913 @code{reg_alloc_order} before execution of the macro.
1915 On most machines, it is not necessary to define this macro.
1918 @node Values in Registers
1919 @subsection How Values Fit in Registers
1921 This section discusses the macros that describe which kinds of values
1922 (specifically, which machine modes) each register can hold, and how many
1923 consecutive registers are needed for a given mode.
1925 @defmac HARD_REGNO_NREGS (@var{regno}, @var{mode})
1926 A C expression for the number of consecutive hard registers, starting
1927 at register number @var{regno}, required to hold a value of mode
1930 On a machine where all registers are exactly one word, a suitable
1931 definition of this macro is
1934 #define HARD_REGNO_NREGS(REGNO, MODE) \
1935 ((GET_MODE_SIZE (MODE) + UNITS_PER_WORD - 1) \
1940 @defmac REGMODE_NATURAL_SIZE (@var{mode})
1941 Define this macro if the natural size of registers that hold values
1942 of mode @var{mode} is not the word size. It is a C expression that
1943 should give the natural size in bytes for the specified mode. It is
1944 used by the register allocator to try to optimize its results. This
1945 happens for example on SPARC 64-bit where the natural size of
1946 floating-point registers is still 32-bit.
1949 @defmac HARD_REGNO_MODE_OK (@var{regno}, @var{mode})
1950 A C expression that is nonzero if it is permissible to store a value
1951 of mode @var{mode} in hard register number @var{regno} (or in several
1952 registers starting with that one). For a machine where all registers
1953 are equivalent, a suitable definition is
1956 #define HARD_REGNO_MODE_OK(REGNO, MODE) 1
1959 You need not include code to check for the numbers of fixed registers,
1960 because the allocation mechanism considers them to be always occupied.
1962 @cindex register pairs
1963 On some machines, double-precision values must be kept in even/odd
1964 register pairs. You can implement that by defining this macro to reject
1965 odd register numbers for such modes.
1967 The minimum requirement for a mode to be OK in a register is that the
1968 @samp{mov@var{mode}} instruction pattern support moves between the
1969 register and other hard register in the same class and that moving a
1970 value into the register and back out not alter it.
1972 Since the same instruction used to move @code{word_mode} will work for
1973 all narrower integer modes, it is not necessary on any machine for
1974 @code{HARD_REGNO_MODE_OK} to distinguish between these modes, provided
1975 you define patterns @samp{movhi}, etc., to take advantage of this. This
1976 is useful because of the interaction between @code{HARD_REGNO_MODE_OK}
1977 and @code{MODES_TIEABLE_P}; it is very desirable for all integer modes
1980 Many machines have special registers for floating point arithmetic.
1981 Often people assume that floating point machine modes are allowed only
1982 in floating point registers. This is not true. Any registers that
1983 can hold integers can safely @emph{hold} a floating point machine
1984 mode, whether or not floating arithmetic can be done on it in those
1985 registers. Integer move instructions can be used to move the values.
1987 On some machines, though, the converse is true: fixed-point machine
1988 modes may not go in floating registers. This is true if the floating
1989 registers normalize any value stored in them, because storing a
1990 non-floating value there would garble it. In this case,
1991 @code{HARD_REGNO_MODE_OK} should reject fixed-point machine modes in
1992 floating registers. But if the floating registers do not automatically
1993 normalize, if you can store any bit pattern in one and retrieve it
1994 unchanged without a trap, then any machine mode may go in a floating
1995 register, so you can define this macro to say so.
1997 The primary significance of special floating registers is rather that
1998 they are the registers acceptable in floating point arithmetic
1999 instructions. However, this is of no concern to
2000 @code{HARD_REGNO_MODE_OK}. You handle it by writing the proper
2001 constraints for those instructions.
2003 On some machines, the floating registers are especially slow to access,
2004 so that it is better to store a value in a stack frame than in such a
2005 register if floating point arithmetic is not being done. As long as the
2006 floating registers are not in class @code{GENERAL_REGS}, they will not
2007 be used unless some pattern's constraint asks for one.
2010 @defmac HARD_REGNO_RENAME_OK (@var{from}, @var{to})
2011 A C expression that is nonzero if it is OK to rename a hard register
2012 @var{from} to another hard register @var{to}.
2014 One common use of this macro is to prevent renaming of a register to
2015 another register that is not saved by a prologue in an interrupt
2018 The default is always nonzero.
2021 @defmac MODES_TIEABLE_P (@var{mode1}, @var{mode2})
2022 A C expression that is nonzero if a value of mode
2023 @var{mode1} is accessible in mode @var{mode2} without copying.
2025 If @code{HARD_REGNO_MODE_OK (@var{r}, @var{mode1})} and
2026 @code{HARD_REGNO_MODE_OK (@var{r}, @var{mode2})} are always the same for
2027 any @var{r}, then @code{MODES_TIEABLE_P (@var{mode1}, @var{mode2})}
2028 should be nonzero. If they differ for any @var{r}, you should define
2029 this macro to return zero unless some other mechanism ensures the
2030 accessibility of the value in a narrower mode.
2032 You should define this macro to return nonzero in as many cases as
2033 possible since doing so will allow GCC to perform better register
2037 @defmac AVOID_CCMODE_COPIES
2038 Define this macro if the compiler should avoid copies to/from @code{CCmode}
2039 registers. You should only define this macro if support for copying to/from
2040 @code{CCmode} is incomplete.
2043 @node Leaf Functions
2044 @subsection Handling Leaf Functions
2046 @cindex leaf functions
2047 @cindex functions, leaf
2048 On some machines, a leaf function (i.e., one which makes no calls) can run
2049 more efficiently if it does not make its own register window. Often this
2050 means it is required to receive its arguments in the registers where they
2051 are passed by the caller, instead of the registers where they would
2054 The special treatment for leaf functions generally applies only when
2055 other conditions are met; for example, often they may use only those
2056 registers for its own variables and temporaries. We use the term ``leaf
2057 function'' to mean a function that is suitable for this special
2058 handling, so that functions with no calls are not necessarily ``leaf
2061 GCC assigns register numbers before it knows whether the function is
2062 suitable for leaf function treatment. So it needs to renumber the
2063 registers in order to output a leaf function. The following macros
2066 @defmac LEAF_REGISTERS
2067 Name of a char vector, indexed by hard register number, which
2068 contains 1 for a register that is allowable in a candidate for leaf
2071 If leaf function treatment involves renumbering the registers, then the
2072 registers marked here should be the ones before renumbering---those that
2073 GCC would ordinarily allocate. The registers which will actually be
2074 used in the assembler code, after renumbering, should not be marked with 1
2077 Define this macro only if the target machine offers a way to optimize
2078 the treatment of leaf functions.
2081 @defmac LEAF_REG_REMAP (@var{regno})
2082 A C expression whose value is the register number to which @var{regno}
2083 should be renumbered, when a function is treated as a leaf function.
2085 If @var{regno} is a register number which should not appear in a leaf
2086 function before renumbering, then the expression should yield @minus{}1, which
2087 will cause the compiler to abort.
2089 Define this macro only if the target machine offers a way to optimize the
2090 treatment of leaf functions, and registers need to be renumbered to do
2094 @findex current_function_is_leaf
2095 @findex current_function_uses_only_leaf_regs
2096 @code{TARGET_ASM_FUNCTION_PROLOGUE} and
2097 @code{TARGET_ASM_FUNCTION_EPILOGUE} must usually treat leaf functions
2098 specially. They can test the C variable @code{current_function_is_leaf}
2099 which is nonzero for leaf functions. @code{current_function_is_leaf} is
2100 set prior to local register allocation and is valid for the remaining
2101 compiler passes. They can also test the C variable
2102 @code{current_function_uses_only_leaf_regs} which is nonzero for leaf
2103 functions which only use leaf registers.
2104 @code{current_function_uses_only_leaf_regs} is valid after all passes
2105 that modify the instructions have been run and is only useful if
2106 @code{LEAF_REGISTERS} is defined.
2107 @c changed this to fix overfull. ALSO: why the "it" at the beginning
2108 @c of the next paragraph?! --mew 2feb93
2110 @node Stack Registers
2111 @subsection Registers That Form a Stack
2113 There are special features to handle computers where some of the
2114 ``registers'' form a stack. Stack registers are normally written by
2115 pushing onto the stack, and are numbered relative to the top of the
2118 Currently, GCC can only handle one group of stack-like registers, and
2119 they must be consecutively numbered. Furthermore, the existing
2120 support for stack-like registers is specific to the 80387 floating
2121 point coprocessor. If you have a new architecture that uses
2122 stack-like registers, you will need to do substantial work on
2123 @file{reg-stack.c} and write your machine description to cooperate
2124 with it, as well as defining these macros.
2127 Define this if the machine has any stack-like registers.
2130 @defmac FIRST_STACK_REG
2131 The number of the first stack-like register. This one is the top
2135 @defmac LAST_STACK_REG
2136 The number of the last stack-like register. This one is the bottom of
2140 @node Register Classes
2141 @section Register Classes
2142 @cindex register class definitions
2143 @cindex class definitions, register
2145 On many machines, the numbered registers are not all equivalent.
2146 For example, certain registers may not be allowed for indexed addressing;
2147 certain registers may not be allowed in some instructions. These machine
2148 restrictions are described to the compiler using @dfn{register classes}.
2150 You define a number of register classes, giving each one a name and saying
2151 which of the registers belong to it. Then you can specify register classes
2152 that are allowed as operands to particular instruction patterns.
2156 In general, each register will belong to several classes. In fact, one
2157 class must be named @code{ALL_REGS} and contain all the registers. Another
2158 class must be named @code{NO_REGS} and contain no registers. Often the
2159 union of two classes will be another class; however, this is not required.
2161 @findex GENERAL_REGS
2162 One of the classes must be named @code{GENERAL_REGS}. There is nothing
2163 terribly special about the name, but the operand constraint letters
2164 @samp{r} and @samp{g} specify this class. If @code{GENERAL_REGS} is
2165 the same as @code{ALL_REGS}, just define it as a macro which expands
2168 Order the classes so that if class @var{x} is contained in class @var{y}
2169 then @var{x} has a lower class number than @var{y}.
2171 The way classes other than @code{GENERAL_REGS} are specified in operand
2172 constraints is through machine-dependent operand constraint letters.
2173 You can define such letters to correspond to various classes, then use
2174 them in operand constraints.
2176 You should define a class for the union of two classes whenever some
2177 instruction allows both classes. For example, if an instruction allows
2178 either a floating point (coprocessor) register or a general register for a
2179 certain operand, you should define a class @code{FLOAT_OR_GENERAL_REGS}
2180 which includes both of them. Otherwise you will get suboptimal code.
2182 You must also specify certain redundant information about the register
2183 classes: for each class, which classes contain it and which ones are
2184 contained in it; for each pair of classes, the largest class contained
2187 When a value occupying several consecutive registers is expected in a
2188 certain class, all the registers used must belong to that class.
2189 Therefore, register classes cannot be used to enforce a requirement for
2190 a register pair to start with an even-numbered register. The way to
2191 specify this requirement is with @code{HARD_REGNO_MODE_OK}.
2193 Register classes used for input-operands of bitwise-and or shift
2194 instructions have a special requirement: each such class must have, for
2195 each fixed-point machine mode, a subclass whose registers can transfer that
2196 mode to or from memory. For example, on some machines, the operations for
2197 single-byte values (@code{QImode}) are limited to certain registers. When
2198 this is so, each register class that is used in a bitwise-and or shift
2199 instruction must have a subclass consisting of registers from which
2200 single-byte values can be loaded or stored. This is so that
2201 @code{PREFERRED_RELOAD_CLASS} can always have a possible value to return.
2203 @deftp {Data type} {enum reg_class}
2204 An enumerated type that must be defined with all the register class names
2205 as enumerated values. @code{NO_REGS} must be first. @code{ALL_REGS}
2206 must be the last register class, followed by one more enumerated value,
2207 @code{LIM_REG_CLASSES}, which is not a register class but rather
2208 tells how many classes there are.
2210 Each register class has a number, which is the value of casting
2211 the class name to type @code{int}. The number serves as an index
2212 in many of the tables described below.
2215 @defmac N_REG_CLASSES
2216 The number of distinct register classes, defined as follows:
2219 #define N_REG_CLASSES (int) LIM_REG_CLASSES
2223 @defmac REG_CLASS_NAMES
2224 An initializer containing the names of the register classes as C string
2225 constants. These names are used in writing some of the debugging dumps.
2228 @defmac REG_CLASS_CONTENTS
2229 An initializer containing the contents of the register classes, as integers
2230 which are bit masks. The @var{n}th integer specifies the contents of class
2231 @var{n}. The way the integer @var{mask} is interpreted is that
2232 register @var{r} is in the class if @code{@var{mask} & (1 << @var{r})} is 1.
2234 When the machine has more than 32 registers, an integer does not suffice.
2235 Then the integers are replaced by sub-initializers, braced groupings containing
2236 several integers. Each sub-initializer must be suitable as an initializer
2237 for the type @code{HARD_REG_SET} which is defined in @file{hard-reg-set.h}.
2238 In this situation, the first integer in each sub-initializer corresponds to
2239 registers 0 through 31, the second integer to registers 32 through 63, and
2243 @defmac REGNO_REG_CLASS (@var{regno})
2244 A C expression whose value is a register class containing hard register
2245 @var{regno}. In general there is more than one such class; choose a class
2246 which is @dfn{minimal}, meaning that no smaller class also contains the
2250 @defmac BASE_REG_CLASS
2251 A macro whose definition is the name of the class to which a valid
2252 base register must belong. A base register is one used in an address
2253 which is the register value plus a displacement.
2256 @defmac MODE_BASE_REG_CLASS (@var{mode})
2257 This is a variation of the @code{BASE_REG_CLASS} macro which allows
2258 the selection of a base register in a mode dependent manner. If
2259 @var{mode} is VOIDmode then it should return the same value as
2260 @code{BASE_REG_CLASS}.
2263 @defmac MODE_BASE_REG_REG_CLASS (@var{mode})
2264 A C expression whose value is the register class to which a valid
2265 base register must belong in order to be used in a base plus index
2266 register address. You should define this macro if base plus index
2267 addresses have different requirements than other base register uses.
2270 @defmac INDEX_REG_CLASS
2271 A macro whose definition is the name of the class to which a valid
2272 index register must belong. An index register is one used in an
2273 address where its value is either multiplied by a scale factor or
2274 added to another register (as well as added to a displacement).
2277 @defmac CONSTRAINT_LEN (@var{char}, @var{str})
2278 For the constraint at the start of @var{str}, which starts with the letter
2279 @var{c}, return the length. This allows you to have register class /
2280 constant / extra constraints that are longer than a single letter;
2281 you don't need to define this macro if you can do with single-letter
2282 constraints only. The definition of this macro should use
2283 DEFAULT_CONSTRAINT_LEN for all the characters that you don't want
2284 to handle specially.
2285 There are some sanity checks in genoutput.c that check the constraint lengths
2286 for the md file, so you can also use this macro to help you while you are
2287 transitioning from a byzantine single-letter-constraint scheme: when you
2288 return a negative length for a constraint you want to re-use, genoutput
2289 will complain about every instance where it is used in the md file.
2292 @defmac REG_CLASS_FROM_LETTER (@var{char})
2293 A C expression which defines the machine-dependent operand constraint
2294 letters for register classes. If @var{char} is such a letter, the
2295 value should be the register class corresponding to it. Otherwise,
2296 the value should be @code{NO_REGS}. The register letter @samp{r},
2297 corresponding to class @code{GENERAL_REGS}, will not be passed
2298 to this macro; you do not need to handle it.
2301 @defmac REG_CLASS_FROM_CONSTRAINT (@var{char}, @var{str})
2302 Like @code{REG_CLASS_FROM_LETTER}, but you also get the constraint string
2303 passed in @var{str}, so that you can use suffixes to distinguish between
2307 @defmac REGNO_OK_FOR_BASE_P (@var{num})
2308 A C expression which is nonzero if register number @var{num} is
2309 suitable for use as a base register in operand addresses. It may be
2310 either a suitable hard register or a pseudo register that has been
2311 allocated such a hard register.
2314 @defmac REGNO_MODE_OK_FOR_BASE_P (@var{num}, @var{mode})
2315 A C expression that is just like @code{REGNO_OK_FOR_BASE_P}, except that
2316 that expression may examine the mode of the memory reference in
2317 @var{mode}. You should define this macro if the mode of the memory
2318 reference affects whether a register may be used as a base register. If
2319 you define this macro, the compiler will use it instead of
2320 @code{REGNO_OK_FOR_BASE_P}.
2323 @defmac REGNO_MODE_OK_FOR_REG_BASE_P (@var{num}, @var{mode})
2324 A C expression which is nonzero if register number @var{num} is suitable for
2325 use as a base register in base plus index operand addresses, accessing
2326 memory in mode @var{mode}. It may be either a suitable hard register or a
2327 pseudo register that has been allocated such a hard register. You should
2328 define this macro if base plus index addresses have different requirements
2329 than other base register uses.
2332 @defmac REGNO_OK_FOR_INDEX_P (@var{num})
2333 A C expression which is nonzero if register number @var{num} is
2334 suitable for use as an index register in operand addresses. It may be
2335 either a suitable hard register or a pseudo register that has been
2336 allocated such a hard register.
2338 The difference between an index register and a base register is that
2339 the index register may be scaled. If an address involves the sum of
2340 two registers, neither one of them scaled, then either one may be
2341 labeled the ``base'' and the other the ``index''; but whichever
2342 labeling is used must fit the machine's constraints of which registers
2343 may serve in each capacity. The compiler will try both labelings,
2344 looking for one that is valid, and will reload one or both registers
2345 only if neither labeling works.
2348 @defmac PREFERRED_RELOAD_CLASS (@var{x}, @var{class})
2349 A C expression that places additional restrictions on the register class
2350 to use when it is necessary to copy value @var{x} into a register in class
2351 @var{class}. The value is a register class; perhaps @var{class}, or perhaps
2352 another, smaller class. On many machines, the following definition is
2356 #define PREFERRED_RELOAD_CLASS(X,CLASS) CLASS
2359 Sometimes returning a more restrictive class makes better code. For
2360 example, on the 68000, when @var{x} is an integer constant that is in range
2361 for a @samp{moveq} instruction, the value of this macro is always
2362 @code{DATA_REGS} as long as @var{class} includes the data registers.
2363 Requiring a data register guarantees that a @samp{moveq} will be used.
2365 One case where @code{PREFERRED_RELOAD_CLASS} must not return
2366 @var{class} is if @var{x} is a legitimate constant which cannot be
2367 loaded into some register class. By returning @code{NO_REGS} you can
2368 force @var{x} into a memory location. For example, rs6000 can load
2369 immediate values into general-purpose registers, but does not have an
2370 instruction for loading an immediate value into a floating-point
2371 register, so @code{PREFERRED_RELOAD_CLASS} returns @code{NO_REGS} when
2372 @var{x} is a floating-point constant. If the constant can't be loaded
2373 into any kind of register, code generation will be better if
2374 @code{LEGITIMATE_CONSTANT_P} makes the constant illegitimate instead
2375 of using @code{PREFERRED_RELOAD_CLASS}.
2378 @defmac PREFERRED_OUTPUT_RELOAD_CLASS (@var{x}, @var{class})
2379 Like @code{PREFERRED_RELOAD_CLASS}, but for output reloads instead of
2380 input reloads. If you don't define this macro, the default is to use
2381 @var{class}, unchanged.
2384 @defmac LIMIT_RELOAD_CLASS (@var{mode}, @var{class})
2385 A C expression that places additional restrictions on the register class
2386 to use when it is necessary to be able to hold a value of mode
2387 @var{mode} in a reload register for which class @var{class} would
2390 Unlike @code{PREFERRED_RELOAD_CLASS}, this macro should be used when
2391 there are certain modes that simply can't go in certain reload classes.
2393 The value is a register class; perhaps @var{class}, or perhaps another,
2396 Don't define this macro unless the target machine has limitations which
2397 require the macro to do something nontrivial.
2400 @deftypefn {Target Hook} enum reg_class TARGET_SECONDARY_RELOAD (bool @var{in_p}, rtx @var{x}, enum reg_class @var{reload_class}, enum machine_mode @var{reload_mode}, secondary_reload_info *@var{sri})
2401 Many machines have some registers that cannot be copied directly to or
2402 from memory or even from other types of registers. An example is the
2403 @samp{MQ} register, which on most machines, can only be copied to or
2404 from general registers, but not memory. Below, we shall be using the
2405 term 'intermediate register' when a move operation cannot be performed
2406 directly, but has to be done by copying the source into the intermediate
2407 register first, and then copying the intermediate register to the
2408 destination. An intermediate register always has the same mode as
2409 source and destination. Since it holds the actual value being copied,
2410 reload might apply optimizations to re-use an intermediate register
2411 and eliding the copy from the source when it can determine that the
2412 intermediate register still holds the required value.
2414 Another kind of secondary reload is required on some machines which
2415 allow copying all registers to and from memory, but require a scratch
2416 register for stores to some memory locations (e.g., those with symbolic
2417 address on the RT, and those with certain symbolic address on the SPARC
2418 when compiling PIC)@. Scratch registers need not have the same mode
2419 as the value being copied, and usually hold a different value that
2420 that being copied. Special patterns in the md file are needed to
2421 describe how the copy is performed with the help of the scratch register;
2422 these patterns also describe the number, register class(es) and mode(s)
2423 of the scratch register(s).
2425 In some cases, both an intermediate and a scratch register are required.
2427 For input reloads, this target hook is called with nonzero @var{in_p},
2428 and @var{x} is an rtx that needs to be copied to a register in of class
2429 @var{reload_class} in @var{reload_mode}. For output reloads, this target
2430 hook is called with zero @var{in_p}, and a register of class @var{reload_mode}
2431 needs to be copied to rtx @var{x} in @var{reload_mode}.
2433 If copying a register of @var{reload_class} from/to @var{x} requires
2434 an intermediate register, the hook @code{secondary_reload} should
2435 return the register class required for this intermediate register.
2436 If no intermediate register is required, it should return NO_REGS.
2437 If more than one intermediate register is required, describe the one
2438 that is closest in the copy chain to the reload register.
2440 If scratch registers are needed, you also have to describe how to
2441 perform the copy from/to the reload register to/from this
2442 closest intermediate register. Or if no intermediate register is
2443 required, but still a scratch register is needed, describe the
2444 copy from/to the reload register to/from the reload operand @var{x}.
2446 You do this by setting @code{sri->icode} to the instruction code of a pattern
2447 in the md file which performs the move. Operands 0 and 1 are the output
2448 and input of this copy, respectively. Operands from operand 2 onward are
2449 for scratch operands. These scratch operands must have a mode, and a
2450 single-register-class
2451 @c [later: or memory]
2454 When an intermediate register is used, the @code{secondary_reload}
2455 hook will be called again to determine how to copy the intermediate
2456 register to/from the reload operand @var{x}, so your hook must also
2457 have code to handle the register class of the intermediate operand.
2459 @c [For later: maybe we'll allow multi-alternative reload patterns -
2460 @c the port maintainer could name a mov<mode> pattern that has clobbers -
2461 @c and match the constraints of input and output to determine the required
2462 @c alternative. A restriction would be that constraints used to match
2463 @c against reloads registers would have to be written as register class
2464 @c constraints, or we need a new target macro / hook that tells us if an
2465 @c arbitrary constraint can match an unknown register of a given class.
2466 @c Such a macro / hook would also be useful in other places.]
2469 @var{x} might be a pseudo-register or a @code{subreg} of a
2470 pseudo-register, which could either be in a hard register or in memory.
2471 Use @code{true_regnum} to find out; it will return @minus{}1 if the pseudo is
2472 in memory and the hard register number if it is in a register.
2474 Scratch operands in memory (constraint @code{"=m"} / @code{"=&m"}) are
2475 currently not supported. For the time being, you will have to continue
2476 to use @code{SECONDARY_MEMORY_NEEDED} for that purpose.
2478 @code{copy_cost} also uses this target hook to find out how values are
2479 copied. If you want it to include some extra cost for the need to allocate
2480 (a) scratch register(s), set @code{sri->extra_cost} to the additional cost.
2481 Or if two dependent moves are supposed to have a lower cost than the sum
2482 of the individual moves due to expected fortuitous scheduling and/or special
2483 forwarding logic, you can set @code{sri->extra_cost} to a negative amount.
2486 @defmac SECONDARY_RELOAD_CLASS (@var{class}, @var{mode}, @var{x})
2487 @defmacx SECONDARY_INPUT_RELOAD_CLASS (@var{class}, @var{mode}, @var{x})
2488 @defmacx SECONDARY_OUTPUT_RELOAD_CLASS (@var{class}, @var{mode}, @var{x})
2489 These macros are obsolete, new ports should use the target hook
2490 @code{TARGET_SECONDARY_RELOAD} instead.
2492 These are obsolete macros, replaced by the @code{TARGET_SECONDARY_RELOAD}
2493 target hook. Older ports still define these macros to indicate to the
2494 reload phase that it may
2495 need to allocate at least one register for a reload in addition to the
2496 register to contain the data. Specifically, if copying @var{x} to a
2497 register @var{class} in @var{mode} requires an intermediate register,
2498 you were supposed to define @code{SECONDARY_INPUT_RELOAD_CLASS} to return the
2499 largest register class all of whose registers can be used as
2500 intermediate registers or scratch registers.
2502 If copying a register @var{class} in @var{mode} to @var{x} requires an
2503 intermediate or scratch register, @code{SECONDARY_OUTPUT_RELOAD_CLASS}
2504 was supposed to be defined be defined to return the largest register
2505 class required. If the
2506 requirements for input and output reloads were the same, the macro
2507 @code{SECONDARY_RELOAD_CLASS} should have been used instead of defining both
2510 The values returned by these macros are often @code{GENERAL_REGS}.
2511 Return @code{NO_REGS} if no spare register is needed; i.e., if @var{x}
2512 can be directly copied to or from a register of @var{class} in
2513 @var{mode} without requiring a scratch register. Do not define this
2514 macro if it would always return @code{NO_REGS}.
2516 If a scratch register is required (either with or without an
2517 intermediate register), you were supposed to define patterns for
2518 @samp{reload_in@var{m}} or @samp{reload_out@var{m}}, as required
2519 (@pxref{Standard Names}. These patterns, which were normally
2520 implemented with a @code{define_expand}, should be similar to the
2521 @samp{mov@var{m}} patterns, except that operand 2 is the scratch
2524 These patterns need constraints for the reload register and scratch
2526 contain a single register class. If the original reload register (whose
2527 class is @var{class}) can meet the constraint given in the pattern, the
2528 value returned by these macros is used for the class of the scratch
2529 register. Otherwise, two additional reload registers are required.
2530 Their classes are obtained from the constraints in the insn pattern.
2532 @var{x} might be a pseudo-register or a @code{subreg} of a
2533 pseudo-register, which could either be in a hard register or in memory.
2534 Use @code{true_regnum} to find out; it will return @minus{}1 if the pseudo is
2535 in memory and the hard register number if it is in a register.
2537 These macros should not be used in the case where a particular class of
2538 registers can only be copied to memory and not to another class of
2539 registers. In that case, secondary reload registers are not needed and
2540 would not be helpful. Instead, a stack location must be used to perform
2541 the copy and the @code{mov@var{m}} pattern should use memory as an
2542 intermediate storage. This case often occurs between floating-point and
2546 @defmac SECONDARY_MEMORY_NEEDED (@var{class1}, @var{class2}, @var{m})
2547 Certain machines have the property that some registers cannot be copied
2548 to some other registers without using memory. Define this macro on
2549 those machines to be a C expression that is nonzero if objects of mode
2550 @var{m} in registers of @var{class1} can only be copied to registers of
2551 class @var{class2} by storing a register of @var{class1} into memory
2552 and loading that memory location into a register of @var{class2}.
2554 Do not define this macro if its value would always be zero.
2557 @defmac SECONDARY_MEMORY_NEEDED_RTX (@var{mode})
2558 Normally when @code{SECONDARY_MEMORY_NEEDED} is defined, the compiler
2559 allocates a stack slot for a memory location needed for register copies.
2560 If this macro is defined, the compiler instead uses the memory location
2561 defined by this macro.
2563 Do not define this macro if you do not define
2564 @code{SECONDARY_MEMORY_NEEDED}.
2567 @defmac SECONDARY_MEMORY_NEEDED_MODE (@var{mode})
2568 When the compiler needs a secondary memory location to copy between two
2569 registers of mode @var{mode}, it normally allocates sufficient memory to
2570 hold a quantity of @code{BITS_PER_WORD} bits and performs the store and
2571 load operations in a mode that many bits wide and whose class is the
2572 same as that of @var{mode}.
2574 This is right thing to do on most machines because it ensures that all
2575 bits of the register are copied and prevents accesses to the registers
2576 in a narrower mode, which some machines prohibit for floating-point
2579 However, this default behavior is not correct on some machines, such as
2580 the DEC Alpha, that store short integers in floating-point registers
2581 differently than in integer registers. On those machines, the default
2582 widening will not work correctly and you must define this macro to
2583 suppress that widening in some cases. See the file @file{alpha.h} for
2586 Do not define this macro if you do not define
2587 @code{SECONDARY_MEMORY_NEEDED} or if widening @var{mode} to a mode that
2588 is @code{BITS_PER_WORD} bits wide is correct for your machine.
2591 @defmac SMALL_REGISTER_CLASSES
2592 On some machines, it is risky to let hard registers live across arbitrary
2593 insns. Typically, these machines have instructions that require values
2594 to be in specific registers (like an accumulator), and reload will fail
2595 if the required hard register is used for another purpose across such an
2598 Define @code{SMALL_REGISTER_CLASSES} to be an expression with a nonzero
2599 value on these machines. When this macro has a nonzero value, the
2600 compiler will try to minimize the lifetime of hard registers.
2602 It is always safe to define this macro with a nonzero value, but if you
2603 unnecessarily define it, you will reduce the amount of optimizations
2604 that can be performed in some cases. If you do not define this macro
2605 with a nonzero value when it is required, the compiler will run out of
2606 spill registers and print a fatal error message. For most machines, you
2607 should not define this macro at all.
2610 @defmac CLASS_LIKELY_SPILLED_P (@var{class})
2611 A C expression whose value is nonzero if pseudos that have been assigned
2612 to registers of class @var{class} would likely be spilled because
2613 registers of @var{class} are needed for spill registers.
2615 The default value of this macro returns 1 if @var{class} has exactly one
2616 register and zero otherwise. On most machines, this default should be
2617 used. Only define this macro to some other expression if pseudos
2618 allocated by @file{local-alloc.c} end up in memory because their hard
2619 registers were needed for spill registers. If this macro returns nonzero
2620 for those classes, those pseudos will only be allocated by
2621 @file{global.c}, which knows how to reallocate the pseudo to another
2622 register. If there would not be another register available for
2623 reallocation, you should not change the definition of this macro since
2624 the only effect of such a definition would be to slow down register
2628 @defmac CLASS_MAX_NREGS (@var{class}, @var{mode})
2629 A C expression for the maximum number of consecutive registers
2630 of class @var{class} needed to hold a value of mode @var{mode}.
2632 This is closely related to the macro @code{HARD_REGNO_NREGS}. In fact,
2633 the value of the macro @code{CLASS_MAX_NREGS (@var{class}, @var{mode})}
2634 should be the maximum value of @code{HARD_REGNO_NREGS (@var{regno},
2635 @var{mode})} for all @var{regno} values in the class @var{class}.
2637 This macro helps control the handling of multiple-word values
2641 @defmac CANNOT_CHANGE_MODE_CLASS (@var{from}, @var{to}, @var{class})
2642 If defined, a C expression that returns nonzero for a @var{class} for which
2643 a change from mode @var{from} to mode @var{to} is invalid.
2645 For the example, loading 32-bit integer or floating-point objects into
2646 floating-point registers on the Alpha extends them to 64 bits.
2647 Therefore loading a 64-bit object and then storing it as a 32-bit object
2648 does not store the low-order 32 bits, as would be the case for a normal
2649 register. Therefore, @file{alpha.h} defines @code{CANNOT_CHANGE_MODE_CLASS}
2653 #define CANNOT_CHANGE_MODE_CLASS(FROM, TO, CLASS) \
2654 (GET_MODE_SIZE (FROM) != GET_MODE_SIZE (TO) \
2655 ? reg_classes_intersect_p (FLOAT_REGS, (CLASS)) : 0)
2659 Three other special macros describe which operands fit which constraint
2662 @defmac CONST_OK_FOR_LETTER_P (@var{value}, @var{c})
2663 A C expression that defines the machine-dependent operand constraint
2664 letters (@samp{I}, @samp{J}, @samp{K}, @dots{} @samp{P}) that specify
2665 particular ranges of integer values. If @var{c} is one of those
2666 letters, the expression should check that @var{value}, an integer, is in
2667 the appropriate range and return 1 if so, 0 otherwise. If @var{c} is
2668 not one of those letters, the value should be 0 regardless of
2672 @defmac CONST_OK_FOR_CONSTRAINT_P (@var{value}, @var{c}, @var{str})
2673 Like @code{CONST_OK_FOR_LETTER_P}, but you also get the constraint
2674 string passed in @var{str}, so that you can use suffixes to distinguish
2675 between different variants.
2678 @defmac CONST_DOUBLE_OK_FOR_LETTER_P (@var{value}, @var{c})
2679 A C expression that defines the machine-dependent operand constraint
2680 letters that specify particular ranges of @code{const_double} values
2681 (@samp{G} or @samp{H}).
2683 If @var{c} is one of those letters, the expression should check that
2684 @var{value}, an RTX of code @code{const_double}, is in the appropriate
2685 range and return 1 if so, 0 otherwise. If @var{c} is not one of those
2686 letters, the value should be 0 regardless of @var{value}.
2688 @code{const_double} is used for all floating-point constants and for
2689 @code{DImode} fixed-point constants. A given letter can accept either
2690 or both kinds of values. It can use @code{GET_MODE} to distinguish
2691 between these kinds.
2694 @defmac CONST_DOUBLE_OK_FOR_CONSTRAINT_P (@var{value}, @var{c}, @var{str})
2695 Like @code{CONST_DOUBLE_OK_FOR_LETTER_P}, but you also get the constraint
2696 string passed in @var{str}, so that you can use suffixes to distinguish
2697 between different variants.
2700 @defmac EXTRA_CONSTRAINT (@var{value}, @var{c})
2701 A C expression that defines the optional machine-dependent constraint
2702 letters that can be used to segregate specific types of operands, usually
2703 memory references, for the target machine. Any letter that is not
2704 elsewhere defined and not matched by @code{REG_CLASS_FROM_LETTER} /
2705 @code{REG_CLASS_FROM_CONSTRAINT}
2706 may be used. Normally this macro will not be defined.
2708 If it is required for a particular target machine, it should return 1
2709 if @var{value} corresponds to the operand type represented by the
2710 constraint letter @var{c}. If @var{c} is not defined as an extra
2711 constraint, the value returned should be 0 regardless of @var{value}.
2713 For example, on the ROMP, load instructions cannot have their output
2714 in r0 if the memory reference contains a symbolic address. Constraint
2715 letter @samp{Q} is defined as representing a memory address that does
2716 @emph{not} contain a symbolic address. An alternative is specified with
2717 a @samp{Q} constraint on the input and @samp{r} on the output. The next
2718 alternative specifies @samp{m} on the input and a register class that
2719 does not include r0 on the output.
2722 @defmac EXTRA_CONSTRAINT_STR (@var{value}, @var{c}, @var{str})
2723 Like @code{EXTRA_CONSTRAINT}, but you also get the constraint string passed
2724 in @var{str}, so that you can use suffixes to distinguish between different
2728 @defmac EXTRA_MEMORY_CONSTRAINT (@var{c}, @var{str})
2729 A C expression that defines the optional machine-dependent constraint
2730 letters, amongst those accepted by @code{EXTRA_CONSTRAINT}, that should
2731 be treated like memory constraints by the reload pass.
2733 It should return 1 if the operand type represented by the constraint
2734 at the start of @var{str}, the first letter of which is the letter @var{c},
2735 comprises a subset of all memory references including
2736 all those whose address is simply a base register. This allows the reload
2737 pass to reload an operand, if it does not directly correspond to the operand
2738 type of @var{c}, by copying its address into a base register.
2740 For example, on the S/390, some instructions do not accept arbitrary
2741 memory references, but only those that do not make use of an index
2742 register. The constraint letter @samp{Q} is defined via
2743 @code{EXTRA_CONSTRAINT} as representing a memory address of this type.
2744 If the letter @samp{Q} is marked as @code{EXTRA_MEMORY_CONSTRAINT},
2745 a @samp{Q} constraint can handle any memory operand, because the
2746 reload pass knows it can be reloaded by copying the memory address
2747 into a base register if required. This is analogous to the way
2748 a @samp{o} constraint can handle any memory operand.
2751 @defmac EXTRA_ADDRESS_CONSTRAINT (@var{c}, @var{str})
2752 A C expression that defines the optional machine-dependent constraint
2753 letters, amongst those accepted by @code{EXTRA_CONSTRAINT} /
2754 @code{EXTRA_CONSTRAINT_STR}, that should
2755 be treated like address constraints by the reload pass.
2757 It should return 1 if the operand type represented by the constraint
2758 at the start of @var{str}, which starts with the letter @var{c}, comprises
2759 a subset of all memory addresses including
2760 all those that consist of just a base register. This allows the reload
2761 pass to reload an operand, if it does not directly correspond to the operand
2762 type of @var{str}, by copying it into a base register.
2764 Any constraint marked as @code{EXTRA_ADDRESS_CONSTRAINT} can only
2765 be used with the @code{address_operand} predicate. It is treated
2766 analogously to the @samp{p} constraint.
2769 @node Stack and Calling
2770 @section Stack Layout and Calling Conventions
2771 @cindex calling conventions
2773 @c prevent bad page break with this line
2774 This describes the stack layout and calling conventions.
2778 * Exception Handling::
2783 * Register Arguments::
2785 * Aggregate Return::
2790 * Stack Smashing Protection::
2794 @subsection Basic Stack Layout
2795 @cindex stack frame layout
2796 @cindex frame layout
2798 @c prevent bad page break with this line
2799 Here is the basic stack layout.
2801 @defmac STACK_GROWS_DOWNWARD
2802 Define this macro if pushing a word onto the stack moves the stack
2803 pointer to a smaller address.
2805 When we say, ``define this macro if @dots{}'', it means that the
2806 compiler checks this macro only with @code{#ifdef} so the precise
2807 definition used does not matter.
2810 @defmac STACK_PUSH_CODE
2811 This macro defines the operation used when something is pushed
2812 on the stack. In RTL, a push operation will be
2813 @code{(set (mem (STACK_PUSH_CODE (reg sp))) @dots{})}
2815 The choices are @code{PRE_DEC}, @code{POST_DEC}, @code{PRE_INC},
2816 and @code{POST_INC}. Which of these is correct depends on
2817 the stack direction and on whether the stack pointer points
2818 to the last item on the stack or whether it points to the
2819 space for the next item on the stack.
2821 The default is @code{PRE_DEC} when @code{STACK_GROWS_DOWNWARD} is
2822 defined, which is almost always right, and @code{PRE_INC} otherwise,
2823 which is often wrong.
2826 @defmac FRAME_GROWS_DOWNWARD
2827 Define this macro to nonzero value if the addresses of local variable slots
2828 are at negative offsets from the frame pointer.
2831 @defmac ARGS_GROW_DOWNWARD
2832 Define this macro if successive arguments to a function occupy decreasing
2833 addresses on the stack.
2836 @defmac STARTING_FRAME_OFFSET
2837 Offset from the frame pointer to the first local variable slot to be allocated.
2839 If @code{FRAME_GROWS_DOWNWARD}, find the next slot's offset by
2840 subtracting the first slot's length from @code{STARTING_FRAME_OFFSET}.
2841 Otherwise, it is found by adding the length of the first slot to the
2842 value @code{STARTING_FRAME_OFFSET}.
2843 @c i'm not sure if the above is still correct.. had to change it to get
2844 @c rid of an overfull. --mew 2feb93
2847 @defmac STACK_ALIGNMENT_NEEDED
2848 Define to zero to disable final alignment of the stack during reload.
2849 The nonzero default for this macro is suitable for most ports.
2851 On ports where @code{STARTING_FRAME_OFFSET} is nonzero or where there
2852 is a register save block following the local block that doesn't require
2853 alignment to @code{STACK_BOUNDARY}, it may be beneficial to disable
2854 stack alignment and do it in the backend.
2857 @defmac STACK_POINTER_OFFSET
2858 Offset from the stack pointer register to the first location at which
2859 outgoing arguments are placed. If not specified, the default value of
2860 zero is used. This is the proper value for most machines.
2862 If @code{ARGS_GROW_DOWNWARD}, this is the offset to the location above
2863 the first location at which outgoing arguments are placed.
2866 @defmac FIRST_PARM_OFFSET (@var{fundecl})
2867 Offset from the argument pointer register to the first argument's
2868 address. On some machines it may depend on the data type of the
2871 If @code{ARGS_GROW_DOWNWARD}, this is the offset to the location above
2872 the first argument's address.
2875 @defmac STACK_DYNAMIC_OFFSET (@var{fundecl})
2876 Offset from the stack pointer register to an item dynamically allocated
2877 on the stack, e.g., by @code{alloca}.
2879 The default value for this macro is @code{STACK_POINTER_OFFSET} plus the
2880 length of the outgoing arguments. The default is correct for most
2881 machines. See @file{function.c} for details.
2884 @defmac INITIAL_FRAME_ADDRESS_RTX
2885 A C expression whose value is RTL representing the address of the initial
2886 stack frame. This address is passed to @code{RETURN_ADDR_RTX} and
2887 @code{DYNAMIC_CHAIN_ADDRESS}. If you don't define this macro, a reasonable
2888 default value will be used. Define this macro in order to make frame pointer
2889 elimination work in the presence of @code{__builtin_frame_address (count)} and
2890 @code{__builtin_return_address (count)} for @code{count} not equal to zero.
2893 @defmac DYNAMIC_CHAIN_ADDRESS (@var{frameaddr})
2894 A C expression whose value is RTL representing the address in a stack
2895 frame where the pointer to the caller's frame is stored. Assume that
2896 @var{frameaddr} is an RTL expression for the address of the stack frame
2899 If you don't define this macro, the default is to return the value
2900 of @var{frameaddr}---that is, the stack frame address is also the
2901 address of the stack word that points to the previous frame.
2904 @defmac SETUP_FRAME_ADDRESSES
2905 If defined, a C expression that produces the machine-specific code to
2906 setup the stack so that arbitrary frames can be accessed. For example,
2907 on the SPARC, we must flush all of the register windows to the stack
2908 before we can access arbitrary stack frames. You will seldom need to
2912 @deftypefn {Target Hook} bool TARGET_BUILTIN_SETJMP_FRAME_VALUE ()
2913 This target hook should return an rtx that is used to store
2914 the address of the current frame into the built in @code{setjmp} buffer.
2915 The default value, @code{virtual_stack_vars_rtx}, is correct for most
2916 machines. One reason you may need to define this target hook is if
2917 @code{hard_frame_pointer_rtx} is the appropriate value on your machine.
2920 @defmac RETURN_ADDR_RTX (@var{count}, @var{frameaddr})
2921 A C expression whose value is RTL representing the value of the return
2922 address for the frame @var{count} steps up from the current frame, after
2923 the prologue. @var{frameaddr} is the frame pointer of the @var{count}
2924 frame, or the frame pointer of the @var{count} @minus{} 1 frame if
2925 @code{RETURN_ADDR_IN_PREVIOUS_FRAME} is defined.
2927 The value of the expression must always be the correct address when
2928 @var{count} is zero, but may be @code{NULL_RTX} if there is not way to
2929 determine the return address of other frames.
2932 @defmac RETURN_ADDR_IN_PREVIOUS_FRAME
2933 Define this if the return address of a particular stack frame is accessed
2934 from the frame pointer of the previous stack frame.
2937 @defmac INCOMING_RETURN_ADDR_RTX
2938 A C expression whose value is RTL representing the location of the
2939 incoming return address at the beginning of any function, before the
2940 prologue. This RTL is either a @code{REG}, indicating that the return
2941 value is saved in @samp{REG}, or a @code{MEM} representing a location in
2944 You only need to define this macro if you want to support call frame
2945 debugging information like that provided by DWARF 2.
2947 If this RTL is a @code{REG}, you should also define
2948 @code{DWARF_FRAME_RETURN_COLUMN} to @code{DWARF_FRAME_REGNUM (REGNO)}.
2951 @defmac DWARF_ALT_FRAME_RETURN_COLUMN
2952 A C expression whose value is an integer giving a DWARF 2 column
2953 number that may be used as an alternate return column. This should
2954 be defined only if @code{DWARF_FRAME_RETURN_COLUMN} is set to a
2955 general register, but an alternate column needs to be used for
2959 @defmac DWARF_ZERO_REG
2960 A C expression whose value is an integer giving a DWARF 2 register
2961 number that is considered to always have the value zero. This should
2962 only be defined if the target has an architected zero register, and
2963 someone decided it was a good idea to use that register number to
2964 terminate the stack backtrace. New ports should avoid this.
2967 @deftypefn {Target Hook} void TARGET_DWARF_HANDLE_FRAME_UNSPEC (const char *@var{label}, rtx @var{pattern}, int @var{index})
2968 This target hook allows the backend to emit frame-related insns that
2969 contain UNSPECs or UNSPEC_VOLATILEs. The DWARF 2 call frame debugging
2970 info engine will invoke it on insns of the form
2972 (set (reg) (unspec [...] UNSPEC_INDEX))
2976 (set (reg) (unspec_volatile [...] UNSPECV_INDEX)).
2978 to let the backend emit the call frame instructions. @var{label} is
2979 the CFI label attached to the insn, @var{pattern} is the pattern of
2980 the insn and @var{index} is @code{UNSPEC_INDEX} or @code{UNSPECV_INDEX}.
2983 @defmac INCOMING_FRAME_SP_OFFSET
2984 A C expression whose value is an integer giving the offset, in bytes,
2985 from the value of the stack pointer register to the top of the stack
2986 frame at the beginning of any function, before the prologue. The top of
2987 the frame is defined to be the value of the stack pointer in the
2988 previous frame, just before the call instruction.
2990 You only need to define this macro if you want to support call frame
2991 debugging information like that provided by DWARF 2.
2994 @defmac ARG_POINTER_CFA_OFFSET (@var{fundecl})
2995 A C expression whose value is an integer giving the offset, in bytes,
2996 from the argument pointer to the canonical frame address (cfa). The
2997 final value should coincide with that calculated by
2998 @code{INCOMING_FRAME_SP_OFFSET}. Which is unfortunately not usable
2999 during virtual register instantiation.
3001 The default value for this macro is @code{FIRST_PARM_OFFSET (fundecl)},
3002 which is correct for most machines; in general, the arguments are found
3003 immediately before the stack frame. Note that this is not the case on
3004 some targets that save registers into the caller's frame, such as SPARC
3005 and rs6000, and so such targets need to define this macro.
3007 You only need to define this macro if the default is incorrect, and you
3008 want to support call frame debugging information like that provided by
3012 @defmac FRAME_POINTER_CFA_OFFSET (@var{fundecl})
3013 If defined, a C expression whose value is an integer giving the offset
3014 in bytes from the frame pointer to the canonical frame address (cfa).
3015 The final value should conincide with that calculated by
3016 @code{INCOMING_FRAME_SP_OFFSET}.
3018 Normally the CFA is calculated as an offset from the argument pointer,
3019 via @code{ARG_POINTER_CFA_OFFSET}, but if the argument pointer is
3020 variable due to the ABI, this may not be possible. If this macro is
3021 defined, it imples that the virtual register instantiation should be
3022 based on the frame pointer instead of the argument pointer. Only one
3023 of @code{FRAME_POINTER_CFA_OFFSET} and @code{ARG_POINTER_CFA_OFFSET}
3027 @node Exception Handling
3028 @subsection Exception Handling Support
3029 @cindex exception handling
3031 @defmac EH_RETURN_DATA_REGNO (@var{N})
3032 A C expression whose value is the @var{N}th register number used for
3033 data by exception handlers, or @code{INVALID_REGNUM} if fewer than
3034 @var{N} registers are usable.
3036 The exception handling library routines communicate with the exception
3037 handlers via a set of agreed upon registers. Ideally these registers
3038 should be call-clobbered; it is possible to use call-saved registers,
3039 but may negatively impact code size. The target must support at least
3040 2 data registers, but should define 4 if there are enough free registers.
3042 You must define this macro if you want to support call frame exception
3043 handling like that provided by DWARF 2.
3046 @defmac EH_RETURN_STACKADJ_RTX
3047 A C expression whose value is RTL representing a location in which
3048 to store a stack adjustment to be applied before function return.
3049 This is used to unwind the stack to an exception handler's call frame.
3050 It will be assigned zero on code paths that return normally.
3052 Typically this is a call-clobbered hard register that is otherwise
3053 untouched by the epilogue, but could also be a stack slot.
3055 Do not define this macro if the stack pointer is saved and restored
3056 by the regular prolog and epilog code in the call frame itself; in
3057 this case, the exception handling library routines will update the
3058 stack location to be restored in place. Otherwise, you must define
3059 this macro if you want to support call frame exception handling like
3060 that provided by DWARF 2.
3063 @defmac EH_RETURN_HANDLER_RTX
3064 A C expression whose value is RTL representing a location in which
3065 to store the address of an exception handler to which we should
3066 return. It will not be assigned on code paths that return normally.
3068 Typically this is the location in the call frame at which the normal
3069 return address is stored. For targets that return by popping an
3070 address off the stack, this might be a memory address just below
3071 the @emph{target} call frame rather than inside the current call
3072 frame. If defined, @code{EH_RETURN_STACKADJ_RTX} will have already
3073 been assigned, so it may be used to calculate the location of the
3076 Some targets have more complex requirements than storing to an
3077 address calculable during initial code generation. In that case
3078 the @code{eh_return} instruction pattern should be used instead.
3080 If you want to support call frame exception handling, you must
3081 define either this macro or the @code{eh_return} instruction pattern.
3084 @defmac RETURN_ADDR_OFFSET
3085 If defined, an integer-valued C expression for which rtl will be generated
3086 to add it to the exception handler address before it is searched in the
3087 exception handling tables, and to subtract it again from the address before
3088 using it to return to the exception handler.
3091 @defmac ASM_PREFERRED_EH_DATA_FORMAT (@var{code}, @var{global})
3092 This macro chooses the encoding of pointers embedded in the exception
3093 handling sections. If at all possible, this should be defined such
3094 that the exception handling section will not require dynamic relocations,
3095 and so may be read-only.
3097 @var{code} is 0 for data, 1 for code labels, 2 for function pointers.
3098 @var{global} is true if the symbol may be affected by dynamic relocations.
3099 The macro should return a combination of the @code{DW_EH_PE_*} defines
3100 as found in @file{dwarf2.h}.
3102 If this macro is not defined, pointers will not be encoded but
3103 represented directly.
3106 @defmac ASM_MAYBE_OUTPUT_ENCODED_ADDR_RTX (@var{file}, @var{encoding}, @var{size}, @var{addr}, @var{done})
3107 This macro allows the target to emit whatever special magic is required
3108 to represent the encoding chosen by @code{ASM_PREFERRED_EH_DATA_FORMAT}.
3109 Generic code takes care of pc-relative and indirect encodings; this must
3110 be defined if the target uses text-relative or data-relative encodings.
3112 This is a C statement that branches to @var{done} if the format was
3113 handled. @var{encoding} is the format chosen, @var{size} is the number
3114 of bytes that the format occupies, @var{addr} is the @code{SYMBOL_REF}
3118 @defmac MD_UNWIND_SUPPORT
3119 A string specifying a file to be #include'd in unwind-dw2.c. The file
3120 so included typically defines @code{MD_FALLBACK_FRAME_STATE_FOR}.
3123 @defmac MD_FALLBACK_FRAME_STATE_FOR (@var{context}, @var{fs})
3124 This macro allows the target to add cpu and operating system specific
3125 code to the call-frame unwinder for use when there is no unwind data
3126 available. The most common reason to implement this macro is to unwind
3127 through signal frames.
3129 This macro is called from @code{uw_frame_state_for} in @file{unwind-dw2.c}
3130 and @file{unwind-ia64.c}. @var{context} is an @code{_Unwind_Context};
3131 @var{fs} is an @code{_Unwind_FrameState}. Examine @code{context->ra}
3132 for the address of the code being executed and @code{context->cfa} for
3133 the stack pointer value. If the frame can be decoded, the register save
3134 addresses should be updated in @var{fs} and the macro should evaluate to
3135 @code{_URC_NO_REASON}. If the frame cannot be decoded, the macro should
3136 evaluate to @code{_URC_END_OF_STACK}.
3138 For proper signal handling in Java this macro is accompanied by
3139 @code{MAKE_THROW_FRAME}, defined in @file{libjava/include/*-signal.h} headers.
3142 @defmac MD_HANDLE_UNWABI (@var{context}, @var{fs})
3143 This macro allows the target to add operating system specific code to the
3144 call-frame unwinder to handle the IA-64 @code{.unwabi} unwinding directive,
3145 usually used for signal or interrupt frames.
3147 This macro is called from @code{uw_update_context} in @file{unwind-ia64.c}.
3148 @var{context} is an @code{_Unwind_Context};
3149 @var{fs} is an @code{_Unwind_FrameState}. Examine @code{fs->unwabi}
3150 for the abi and context in the @code{.unwabi} directive. If the
3151 @code{.unwabi} directive can be handled, the register save addresses should
3152 be updated in @var{fs}.
3155 @defmac TARGET_USES_WEAK_UNWIND_INFO
3156 A C expression that evaluates to true if the target requires unwind
3157 info to be given comdat linkage. Define it to be @code{1} if comdat
3158 linkage is necessary. The default is @code{0}.
3161 @node Stack Checking
3162 @subsection Specifying How Stack Checking is Done
3164 GCC will check that stack references are within the boundaries of
3165 the stack, if the @option{-fstack-check} is specified, in one of three ways:
3169 If the value of the @code{STACK_CHECK_BUILTIN} macro is nonzero, GCC
3170 will assume that you have arranged for stack checking to be done at
3171 appropriate places in the configuration files, e.g., in
3172 @code{TARGET_ASM_FUNCTION_PROLOGUE}. GCC will do not other special
3176 If @code{STACK_CHECK_BUILTIN} is zero and you defined a named pattern
3177 called @code{check_stack} in your @file{md} file, GCC will call that
3178 pattern with one argument which is the address to compare the stack
3179 value against. You must arrange for this pattern to report an error if
3180 the stack pointer is out of range.
3183 If neither of the above are true, GCC will generate code to periodically
3184 ``probe'' the stack pointer using the values of the macros defined below.
3187 Normally, you will use the default values of these macros, so GCC
3188 will use the third approach.
3190 @defmac STACK_CHECK_BUILTIN
3191 A nonzero value if stack checking is done by the configuration files in a
3192 machine-dependent manner. You should define this macro if stack checking
3193 is require by the ABI of your machine or if you would like to have to stack
3194 checking in some more efficient way than GCC's portable approach.
3195 The default value of this macro is zero.
3198 @defmac STACK_CHECK_PROBE_INTERVAL
3199 An integer representing the interval at which GCC must generate stack
3200 probe instructions. You will normally define this macro to be no larger
3201 than the size of the ``guard pages'' at the end of a stack area. The
3202 default value of 4096 is suitable for most systems.
3205 @defmac STACK_CHECK_PROBE_LOAD
3206 A integer which is nonzero if GCC should perform the stack probe
3207 as a load instruction and zero if GCC should use a store instruction.
3208 The default is zero, which is the most efficient choice on most systems.
3211 @defmac STACK_CHECK_PROTECT
3212 The number of bytes of stack needed to recover from a stack overflow,
3213 for languages where such a recovery is supported. The default value of
3214 75 words should be adequate for most machines.
3217 @defmac STACK_CHECK_MAX_FRAME_SIZE
3218 The maximum size of a stack frame, in bytes. GCC will generate probe
3219 instructions in non-leaf functions to ensure at least this many bytes of
3220 stack are available. If a stack frame is larger than this size, stack
3221 checking will not be reliable and GCC will issue a warning. The
3222 default is chosen so that GCC only generates one instruction on most
3223 systems. You should normally not change the default value of this macro.
3226 @defmac STACK_CHECK_FIXED_FRAME_SIZE
3227 GCC uses this value to generate the above warning message. It
3228 represents the amount of fixed frame used by a function, not including
3229 space for any callee-saved registers, temporaries and user variables.
3230 You need only specify an upper bound for this amount and will normally
3231 use the default of four words.
3234 @defmac STACK_CHECK_MAX_VAR_SIZE
3235 The maximum size, in bytes, of an object that GCC will place in the
3236 fixed area of the stack frame when the user specifies
3237 @option{-fstack-check}.
3238 GCC computed the default from the values of the above macros and you will
3239 normally not need to override that default.
3243 @node Frame Registers
3244 @subsection Registers That Address the Stack Frame
3246 @c prevent bad page break with this line
3247 This discusses registers that address the stack frame.
3249 @defmac STACK_POINTER_REGNUM
3250 The register number of the stack pointer register, which must also be a
3251 fixed register according to @code{FIXED_REGISTERS}. On most machines,
3252 the hardware determines which register this is.
3255 @defmac FRAME_POINTER_REGNUM
3256 The register number of the frame pointer register, which is used to
3257 access automatic variables in the stack frame. On some machines, the
3258 hardware determines which register this is. On other machines, you can
3259 choose any register you wish for this purpose.
3262 @defmac HARD_FRAME_POINTER_REGNUM
3263 On some machines the offset between the frame pointer and starting
3264 offset of the automatic variables is not known until after register
3265 allocation has been done (for example, because the saved registers are
3266 between these two locations). On those machines, define
3267 @code{FRAME_POINTER_REGNUM} the number of a special, fixed register to
3268 be used internally until the offset is known, and define
3269 @code{HARD_FRAME_POINTER_REGNUM} to be the actual hard register number
3270 used for the frame pointer.
3272 You should define this macro only in the very rare circumstances when it
3273 is not possible to calculate the offset between the frame pointer and
3274 the automatic variables until after register allocation has been
3275 completed. When this macro is defined, you must also indicate in your
3276 definition of @code{ELIMINABLE_REGS} how to eliminate
3277 @code{FRAME_POINTER_REGNUM} into either @code{HARD_FRAME_POINTER_REGNUM}
3278 or @code{STACK_POINTER_REGNUM}.
3280 Do not define this macro if it would be the same as
3281 @code{FRAME_POINTER_REGNUM}.
3284 @defmac ARG_POINTER_REGNUM
3285 The register number of the arg pointer register, which is used to access
3286 the function's argument list. On some machines, this is the same as the
3287 frame pointer register. On some machines, the hardware determines which
3288 register this is. On other machines, you can choose any register you
3289 wish for this purpose. If this is not the same register as the frame
3290 pointer register, then you must mark it as a fixed register according to
3291 @code{FIXED_REGISTERS}, or arrange to be able to eliminate it
3292 (@pxref{Elimination}).
3295 @defmac RETURN_ADDRESS_POINTER_REGNUM
3296 The register number of the return address pointer register, which is used to
3297 access the current function's return address from the stack. On some
3298 machines, the return address is not at a fixed offset from the frame
3299 pointer or stack pointer or argument pointer. This register can be defined
3300 to point to the return address on the stack, and then be converted by
3301 @code{ELIMINABLE_REGS} into either the frame pointer or stack pointer.
3303 Do not define this macro unless there is no other way to get the return
3304 address from the stack.
3307 @defmac STATIC_CHAIN_REGNUM
3308 @defmacx STATIC_CHAIN_INCOMING_REGNUM
3309 Register numbers used for passing a function's static chain pointer. If
3310 register windows are used, the register number as seen by the called
3311 function is @code{STATIC_CHAIN_INCOMING_REGNUM}, while the register
3312 number as seen by the calling function is @code{STATIC_CHAIN_REGNUM}. If
3313 these registers are the same, @code{STATIC_CHAIN_INCOMING_REGNUM} need
3316 The static chain register need not be a fixed register.
3318 If the static chain is passed in memory, these macros should not be
3319 defined; instead, the next two macros should be defined.
3322 @defmac STATIC_CHAIN
3323 @defmacx STATIC_CHAIN_INCOMING
3324 If the static chain is passed in memory, these macros provide rtx giving
3325 @code{mem} expressions that denote where they are stored.
3326 @code{STATIC_CHAIN} and @code{STATIC_CHAIN_INCOMING} give the locations
3327 as seen by the calling and called functions, respectively. Often the former
3328 will be at an offset from the stack pointer and the latter at an offset from
3331 @findex stack_pointer_rtx
3332 @findex frame_pointer_rtx
3333 @findex arg_pointer_rtx
3334 The variables @code{stack_pointer_rtx}, @code{frame_pointer_rtx}, and
3335 @code{arg_pointer_rtx} will have been initialized prior to the use of these
3336 macros and should be used to refer to those items.
3338 If the static chain is passed in a register, the two previous macros should
3342 @defmac DWARF_FRAME_REGISTERS
3343 This macro specifies the maximum number of hard registers that can be
3344 saved in a call frame. This is used to size data structures used in
3345 DWARF2 exception handling.
3347 Prior to GCC 3.0, this macro was needed in order to establish a stable
3348 exception handling ABI in the face of adding new hard registers for ISA
3349 extensions. In GCC 3.0 and later, the EH ABI is insulated from changes
3350 in the number of hard registers. Nevertheless, this macro can still be
3351 used to reduce the runtime memory requirements of the exception handling
3352 routines, which can be substantial if the ISA contains a lot of
3353 registers that are not call-saved.
3355 If this macro is not defined, it defaults to
3356 @code{FIRST_PSEUDO_REGISTER}.
3359 @defmac PRE_GCC3_DWARF_FRAME_REGISTERS
3361 This macro is similar to @code{DWARF_FRAME_REGISTERS}, but is provided
3362 for backward compatibility in pre GCC 3.0 compiled code.
3364 If this macro is not defined, it defaults to
3365 @code{DWARF_FRAME_REGISTERS}.
3368 @defmac DWARF_REG_TO_UNWIND_COLUMN (@var{regno})
3370 Define this macro if the target's representation for dwarf registers
3371 is different than the internal representation for unwind column.
3372 Given a dwarf register, this macro should return the internal unwind
3373 column number to use instead.
3375 See the PowerPC's SPE target for an example.
3378 @defmac DWARF_FRAME_REGNUM (@var{regno})
3380 Define this macro if the target's representation for dwarf registers
3381 used in .eh_frame or .debug_frame is different from that used in other
3382 debug info sections. Given a GCC hard register number, this macro
3383 should return the .eh_frame register number. The default is
3384 @code{DBX_REGISTER_NUMBER (@var{regno})}.
3388 @defmac DWARF2_FRAME_REG_OUT (@var{regno}, @var{for_eh})
3390 Define this macro to map register numbers held in the call frame info
3391 that GCC has collected using @code{DWARF_FRAME_REGNUM} to those that
3392 should be output in .debug_frame (@code{@var{for_eh}} is zero) and
3393 .eh_frame (@code{@var{for_eh}} is nonzero). The default is to
3394 return @code{@var{regno}}.
3399 @subsection Eliminating Frame Pointer and Arg Pointer
3401 @c prevent bad page break with this line
3402 This is about eliminating the frame pointer and arg pointer.
3404 @defmac FRAME_POINTER_REQUIRED
3405 A C expression which is nonzero if a function must have and use a frame
3406 pointer. This expression is evaluated in the reload pass. If its value is
3407 nonzero the function will have a frame pointer.
3409 The expression can in principle examine the current function and decide
3410 according to the facts, but on most machines the constant 0 or the
3411 constant 1 suffices. Use 0 when the machine allows code to be generated
3412 with no frame pointer, and doing so saves some time or space. Use 1
3413 when there is no possible advantage to avoiding a frame pointer.
3415 In certain cases, the compiler does not know how to produce valid code
3416 without a frame pointer. The compiler recognizes those cases and
3417 automatically gives the function a frame pointer regardless of what
3418 @code{FRAME_POINTER_REQUIRED} says. You don't need to worry about
3421 In a function that does not require a frame pointer, the frame pointer
3422 register can be allocated for ordinary usage, unless you mark it as a
3423 fixed register. See @code{FIXED_REGISTERS} for more information.
3426 @findex get_frame_size
3427 @defmac INITIAL_FRAME_POINTER_OFFSET (@var{depth-var})
3428 A C statement to store in the variable @var{depth-var} the difference
3429 between the frame pointer and the stack pointer values immediately after
3430 the function prologue. The value would be computed from information
3431 such as the result of @code{get_frame_size ()} and the tables of
3432 registers @code{regs_ever_live} and @code{call_used_regs}.
3434 If @code{ELIMINABLE_REGS} is defined, this macro will be not be used and
3435 need not be defined. Otherwise, it must be defined even if
3436 @code{FRAME_POINTER_REQUIRED} is defined to always be true; in that
3437 case, you may set @var{depth-var} to anything.
3440 @defmac ELIMINABLE_REGS
3441 If defined, this macro specifies a table of register pairs used to
3442 eliminate unneeded registers that point into the stack frame. If it is not
3443 defined, the only elimination attempted by the compiler is to replace
3444 references to the frame pointer with references to the stack pointer.
3446 The definition of this macro is a list of structure initializations, each
3447 of which specifies an original and replacement register.
3449 On some machines, the position of the argument pointer is not known until
3450 the compilation is completed. In such a case, a separate hard register
3451 must be used for the argument pointer. This register can be eliminated by
3452 replacing it with either the frame pointer or the argument pointer,
3453 depending on whether or not the frame pointer has been eliminated.
3455 In this case, you might specify:
3457 #define ELIMINABLE_REGS \
3458 @{@{ARG_POINTER_REGNUM, STACK_POINTER_REGNUM@}, \
3459 @{ARG_POINTER_REGNUM, FRAME_POINTER_REGNUM@}, \
3460 @{FRAME_POINTER_REGNUM, STACK_POINTER_REGNUM@}@}
3463 Note that the elimination of the argument pointer with the stack pointer is
3464 specified first since that is the preferred elimination.
3467 @defmac CAN_ELIMINATE (@var{from-reg}, @var{to-reg})
3468 A C expression that returns nonzero if the compiler is allowed to try
3469 to replace register number @var{from-reg} with register number
3470 @var{to-reg}. This macro need only be defined if @code{ELIMINABLE_REGS}
3471 is defined, and will usually be the constant 1, since most of the cases
3472 preventing register elimination are things that the compiler already
3476 @defmac INITIAL_ELIMINATION_OFFSET (@var{from-reg}, @var{to-reg}, @var{offset-var})
3477 This macro is similar to @code{INITIAL_FRAME_POINTER_OFFSET}. It
3478 specifies the initial difference between the specified pair of
3479 registers. This macro must be defined if @code{ELIMINABLE_REGS} is
3483 @node Stack Arguments
3484 @subsection Passing Function Arguments on the Stack
3485 @cindex arguments on stack
3486 @cindex stack arguments
3488 The macros in this section control how arguments are passed
3489 on the stack. See the following section for other macros that
3490 control passing certain arguments in registers.
3492 @deftypefn {Target Hook} bool TARGET_PROMOTE_PROTOTYPES (tree @var{fntype})
3493 This target hook returns @code{true} if an argument declared in a
3494 prototype as an integral type smaller than @code{int} should actually be
3495 passed as an @code{int}. In addition to avoiding errors in certain
3496 cases of mismatch, it also makes for better code on certain machines.
3497 The default is to not promote prototypes.
3501 A C expression. If nonzero, push insns will be used to pass
3503 If the target machine does not have a push instruction, set it to zero.
3504 That directs GCC to use an alternate strategy: to
3505 allocate the entire argument block and then store the arguments into
3506 it. When @code{PUSH_ARGS} is nonzero, @code{PUSH_ROUNDING} must be defined too.
3509 @defmac PUSH_ARGS_REVERSED
3510 A C expression. If nonzero, function arguments will be evaluated from
3511 last to first, rather than from first to last. If this macro is not
3512 defined, it defaults to @code{PUSH_ARGS} on targets where the stack
3513 and args grow in opposite directions, and 0 otherwise.
3516 @defmac PUSH_ROUNDING (@var{npushed})
3517 A C expression that is the number of bytes actually pushed onto the
3518 stack when an instruction attempts to push @var{npushed} bytes.
3520 On some machines, the definition
3523 #define PUSH_ROUNDING(BYTES) (BYTES)
3527 will suffice. But on other machines, instructions that appear
3528 to push one byte actually push two bytes in an attempt to maintain
3529 alignment. Then the definition should be
3532 #define PUSH_ROUNDING(BYTES) (((BYTES) + 1) & ~1)
3536 @findex current_function_outgoing_args_size
3537 @defmac ACCUMULATE_OUTGOING_ARGS
3538 A C expression. If nonzero, the maximum amount of space required for outgoing arguments
3539 will be computed and placed into the variable
3540 @code{current_function_outgoing_args_size}. No space will be pushed
3541 onto the stack for each call; instead, the function prologue should
3542 increase the stack frame size by this amount.
3544 Setting both @code{PUSH_ARGS} and @code{ACCUMULATE_OUTGOING_ARGS}
3548 @defmac REG_PARM_STACK_SPACE (@var{fndecl})
3549 Define this macro if functions should assume that stack space has been
3550 allocated for arguments even when their values are passed in
3553 The value of this macro is the size, in bytes, of the area reserved for
3554 arguments passed in registers for the function represented by @var{fndecl},
3555 which can be zero if GCC is calling a library function.
3557 This space can be allocated by the caller, or be a part of the
3558 machine-dependent stack frame: @code{OUTGOING_REG_PARM_STACK_SPACE} says
3561 @c above is overfull. not sure what to do. --mew 5feb93 did
3562 @c something, not sure if it looks good. --mew 10feb93
3564 @defmac OUTGOING_REG_PARM_STACK_SPACE
3565 Define this if it is the responsibility of the caller to allocate the area
3566 reserved for arguments passed in registers.
3568 If @code{ACCUMULATE_OUTGOING_ARGS} is defined, this macro controls
3569 whether the space for these arguments counts in the value of
3570 @code{current_function_outgoing_args_size}.
3573 @defmac STACK_PARMS_IN_REG_PARM_AREA
3574 Define this macro if @code{REG_PARM_STACK_SPACE} is defined, but the
3575 stack parameters don't skip the area specified by it.
3576 @c i changed this, makes more sens and it should have taken care of the
3577 @c overfull.. not as specific, tho. --mew 5feb93
3579 Normally, when a parameter is not passed in registers, it is placed on the
3580 stack beyond the @code{REG_PARM_STACK_SPACE} area. Defining this macro
3581 suppresses this behavior and causes the parameter to be passed on the
3582 stack in its natural location.
3585 @defmac RETURN_POPS_ARGS (@var{fundecl}, @var{funtype}, @var{stack-size})
3586 A C expression that should indicate the number of bytes of its own
3587 arguments that a function pops on returning, or 0 if the
3588 function pops no arguments and the caller must therefore pop them all
3589 after the function returns.
3591 @var{fundecl} is a C variable whose value is a tree node that describes
3592 the function in question. Normally it is a node of type
3593 @code{FUNCTION_DECL} that describes the declaration of the function.
3594 From this you can obtain the @code{DECL_ATTRIBUTES} of the function.
3596 @var{funtype} is a C variable whose value is a tree node that
3597 describes the function in question. Normally it is a node of type
3598 @code{FUNCTION_TYPE} that describes the data type of the function.
3599 From this it is possible to obtain the data types of the value and
3600 arguments (if known).
3602 When a call to a library function is being considered, @var{fundecl}
3603 will contain an identifier node for the library function. Thus, if
3604 you need to distinguish among various library functions, you can do so
3605 by their names. Note that ``library function'' in this context means
3606 a function used to perform arithmetic, whose name is known specially
3607 in the compiler and was not mentioned in the C code being compiled.
3609 @var{stack-size} is the number of bytes of arguments passed on the
3610 stack. If a variable number of bytes is passed, it is zero, and
3611 argument popping will always be the responsibility of the calling function.
3613 On the VAX, all functions always pop their arguments, so the definition
3614 of this macro is @var{stack-size}. On the 68000, using the standard
3615 calling convention, no functions pop their arguments, so the value of
3616 the macro is always 0 in this case. But an alternative calling
3617 convention is available in which functions that take a fixed number of
3618 arguments pop them but other functions (such as @code{printf}) pop
3619 nothing (the caller pops all). When this convention is in use,
3620 @var{funtype} is examined to determine whether a function takes a fixed
3621 number of arguments.
3624 @defmac CALL_POPS_ARGS (@var{cum})
3625 A C expression that should indicate the number of bytes a call sequence
3626 pops off the stack. It is added to the value of @code{RETURN_POPS_ARGS}
3627 when compiling a function call.
3629 @var{cum} is the variable in which all arguments to the called function
3630 have been accumulated.
3632 On certain architectures, such as the SH5, a call trampoline is used
3633 that pops certain registers off the stack, depending on the arguments
3634 that have been passed to the function. Since this is a property of the
3635 call site, not of the called function, @code{RETURN_POPS_ARGS} is not
3639 @node Register Arguments
3640 @subsection Passing Arguments in Registers
3641 @cindex arguments in registers
3642 @cindex registers arguments
3644 This section describes the macros which let you control how various
3645 types of arguments are passed in registers or how they are arranged in
3648 @defmac FUNCTION_ARG (@var{cum}, @var{mode}, @var{type}, @var{named})
3649 A C expression that controls whether a function argument is passed
3650 in a register, and which register.
3652 The arguments are @var{cum}, which summarizes all the previous
3653 arguments; @var{mode}, the machine mode of the argument; @var{type},
3654 the data type of the argument as a tree node or 0 if that is not known
3655 (which happens for C support library functions); and @var{named},
3656 which is 1 for an ordinary argument and 0 for nameless arguments that
3657 correspond to @samp{@dots{}} in the called function's prototype.
3658 @var{type} can be an incomplete type if a syntax error has previously
3661 The value of the expression is usually either a @code{reg} RTX for the
3662 hard register in which to pass the argument, or zero to pass the
3663 argument on the stack.
3665 For machines like the VAX and 68000, where normally all arguments are
3666 pushed, zero suffices as a definition.
3668 The value of the expression can also be a @code{parallel} RTX@. This is
3669 used when an argument is passed in multiple locations. The mode of the
3670 @code{parallel} should be the mode of the entire argument. The
3671 @code{parallel} holds any number of @code{expr_list} pairs; each one
3672 describes where part of the argument is passed. In each
3673 @code{expr_list} the first operand must be a @code{reg} RTX for the hard
3674 register in which to pass this part of the argument, and the mode of the
3675 register RTX indicates how large this part of the argument is. The
3676 second operand of the @code{expr_list} is a @code{const_int} which gives
3677 the offset in bytes into the entire argument of where this part starts.
3678 As a special exception the first @code{expr_list} in the @code{parallel}
3679 RTX may have a first operand of zero. This indicates that the entire
3680 argument is also stored on the stack.
3682 The last time this macro is called, it is called with @code{MODE ==
3683 VOIDmode}, and its result is passed to the @code{call} or @code{call_value}
3684 pattern as operands 2 and 3 respectively.
3686 @cindex @file{stdarg.h} and register arguments
3687 The usual way to make the ISO library @file{stdarg.h} work on a machine
3688 where some arguments are usually passed in registers, is to cause
3689 nameless arguments to be passed on the stack instead. This is done
3690 by making @code{FUNCTION_ARG} return 0 whenever @var{named} is 0.
3692 @cindex @code{TARGET_MUST_PASS_IN_STACK}, and @code{FUNCTION_ARG}
3693 @cindex @code{REG_PARM_STACK_SPACE}, and @code{FUNCTION_ARG}
3694 You may use the hook @code{targetm.calls.must_pass_in_stack}
3695 in the definition of this macro to determine if this argument is of a
3696 type that must be passed in the stack. If @code{REG_PARM_STACK_SPACE}
3697 is not defined and @code{FUNCTION_ARG} returns nonzero for such an
3698 argument, the compiler will abort. If @code{REG_PARM_STACK_SPACE} is
3699 defined, the argument will be computed in the stack and then loaded into
3703 @deftypefn {Target Hook} bool TARGET_MUST_PASS_IN_STACK (enum machine_mode @var{mode}, tree @var{type})
3704 This target hook should return @code{true} if we should not pass @var{type}
3705 solely in registers. The file @file{expr.h} defines a
3706 definition that is usually appropriate, refer to @file{expr.h} for additional
3710 @defmac FUNCTION_INCOMING_ARG (@var{cum}, @var{mode}, @var{type}, @var{named})
3711 Define this macro if the target machine has ``register windows'', so
3712 that the register in which a function sees an arguments is not
3713 necessarily the same as the one in which the caller passed the
3716 For such machines, @code{FUNCTION_ARG} computes the register in which
3717 the caller passes the value, and @code{FUNCTION_INCOMING_ARG} should
3718 be defined in a similar fashion to tell the function being called
3719 where the arguments will arrive.
3721 If @code{FUNCTION_INCOMING_ARG} is not defined, @code{FUNCTION_ARG}
3722 serves both purposes.
3725 @deftypefn {Target Hook} int TARGET_ARG_PARTIAL_BYTES (CUMULATIVE_ARGS *@var{cum}, enum machine_mode @var{mode}, tree @var{type}, bool @var{named})
3726 This target hook returns the number of bytes at the beginning of an
3727 argument that must be put in registers. The value must be zero for
3728 arguments that are passed entirely in registers or that are entirely
3729 pushed on the stack.
3731 On some machines, certain arguments must be passed partially in
3732 registers and partially in memory. On these machines, typically the
3733 first few words of arguments are passed in registers, and the rest
3734 on the stack. If a multi-word argument (a @code{double} or a
3735 structure) crosses that boundary, its first few words must be passed
3736 in registers and the rest must be pushed. This macro tells the
3737 compiler when this occurs, and how many bytes should go in registers.
3739 @code{FUNCTION_ARG} for these arguments should return the first
3740 register to be used by the caller for this argument; likewise
3741 @code{FUNCTION_INCOMING_ARG}, for the called function.
3744 @deftypefn {Target Hook} bool TARGET_PASS_BY_REFERENCE (CUMULATIVE_ARGS *@var{cum}, enum machine_mode @var{mode}, tree @var{type}, bool @var{named})
3745 This target hook should return @code{true} if an argument at the
3746 position indicated by @var{cum} should be passed by reference. This
3747 predicate is queried after target independent reasons for being
3748 passed by reference, such as @code{TREE_ADDRESSABLE (type)}.
3750 If the hook returns true, a copy of that argument is made in memory and a
3751 pointer to the argument is passed instead of the argument itself.
3752 The pointer is passed in whatever way is appropriate for passing a pointer
3756 @deftypefn {Target Hook} bool TARGET_CALLEE_COPIES (CUMULATIVE_ARGS *@var{cum}, enum machine_mode @var{mode}, tree @var{type}, bool @var{named})
3757 The function argument described by the parameters to this hook is
3758 known to be passed by reference. The hook should return true if the
3759 function argument should be copied by the callee instead of copied
3762 For any argument for which the hook returns true, if it can be
3763 determined that the argument is not modified, then a copy need
3766 The default version of this hook always returns false.
3769 @defmac CUMULATIVE_ARGS
3770 A C type for declaring a variable that is used as the first argument of
3771 @code{FUNCTION_ARG} and other related values. For some target machines,
3772 the type @code{int} suffices and can hold the number of bytes of
3775 There is no need to record in @code{CUMULATIVE_ARGS} anything about the
3776 arguments that have been passed on the stack. The compiler has other
3777 variables to keep track of that. For target machines on which all
3778 arguments are passed on the stack, there is no need to store anything in
3779 @code{CUMULATIVE_ARGS}; however, the data structure must exist and
3780 should not be empty, so use @code{int}.
3783 @defmac INIT_CUMULATIVE_ARGS (@var{cum}, @var{fntype}, @var{libname}, @var{fndecl}, @var{n_named_args})
3784 A C statement (sans semicolon) for initializing the variable
3785 @var{cum} for the state at the beginning of the argument list. The
3786 variable has type @code{CUMULATIVE_ARGS}. The value of @var{fntype}
3787 is the tree node for the data type of the function which will receive
3788 the args, or 0 if the args are to a compiler support library function.
3789 For direct calls that are not libcalls, @var{fndecl} contain the
3790 declaration node of the function. @var{fndecl} is also set when
3791 @code{INIT_CUMULATIVE_ARGS} is used to find arguments for the function
3792 being compiled. @var{n_named_args} is set to the number of named
3793 arguments, including a structure return address if it is passed as a
3794 parameter, when making a call. When processing incoming arguments,
3795 @var{n_named_args} is set to @minus{}1.
3797 When processing a call to a compiler support library function,
3798 @var{libname} identifies which one. It is a @code{symbol_ref} rtx which
3799 contains the name of the function, as a string. @var{libname} is 0 when
3800 an ordinary C function call is being processed. Thus, each time this
3801 macro is called, either @var{libname} or @var{fntype} is nonzero, but
3802 never both of them at once.
3805 @defmac INIT_CUMULATIVE_LIBCALL_ARGS (@var{cum}, @var{mode}, @var{libname})
3806 Like @code{INIT_CUMULATIVE_ARGS} but only used for outgoing libcalls,
3807 it gets a @code{MODE} argument instead of @var{fntype}, that would be
3808 @code{NULL}. @var{indirect} would always be zero, too. If this macro
3809 is not defined, @code{INIT_CUMULATIVE_ARGS (cum, NULL_RTX, libname,
3810 0)} is used instead.
3813 @defmac INIT_CUMULATIVE_INCOMING_ARGS (@var{cum}, @var{fntype}, @var{libname})
3814 Like @code{INIT_CUMULATIVE_ARGS} but overrides it for the purposes of
3815 finding the arguments for the function being compiled. If this macro is
3816 undefined, @code{INIT_CUMULATIVE_ARGS} is used instead.
3818 The value passed for @var{libname} is always 0, since library routines
3819 with special calling conventions are never compiled with GCC@. The
3820 argument @var{libname} exists for symmetry with
3821 @code{INIT_CUMULATIVE_ARGS}.
3822 @c could use "this macro" in place of @code{INIT_CUMULATIVE_ARGS}, maybe.
3823 @c --mew 5feb93 i switched the order of the sentences. --mew 10feb93
3826 @defmac FUNCTION_ARG_ADVANCE (@var{cum}, @var{mode}, @var{type}, @var{named})
3827 A C statement (sans semicolon) to update the summarizer variable
3828 @var{cum} to advance past an argument in the argument list. The
3829 values @var{mode}, @var{type} and @var{named} describe that argument.
3830 Once this is done, the variable @var{cum} is suitable for analyzing
3831 the @emph{following} argument with @code{FUNCTION_ARG}, etc.
3833 This macro need not do anything if the argument in question was passed
3834 on the stack. The compiler knows how to track the amount of stack space
3835 used for arguments without any special help.
3838 @defmac FUNCTION_ARG_PADDING (@var{mode}, @var{type})
3839 If defined, a C expression which determines whether, and in which direction,
3840 to pad out an argument with extra space. The value should be of type
3841 @code{enum direction}: either @code{upward} to pad above the argument,
3842 @code{downward} to pad below, or @code{none} to inhibit padding.
3844 The @emph{amount} of padding is always just enough to reach the next
3845 multiple of @code{FUNCTION_ARG_BOUNDARY}; this macro does not control
3848 This macro has a default definition which is right for most systems.
3849 For little-endian machines, the default is to pad upward. For
3850 big-endian machines, the default is to pad downward for an argument of
3851 constant size shorter than an @code{int}, and upward otherwise.
3854 @defmac PAD_VARARGS_DOWN
3855 If defined, a C expression which determines whether the default
3856 implementation of va_arg will attempt to pad down before reading the
3857 next argument, if that argument is smaller than its aligned space as
3858 controlled by @code{PARM_BOUNDARY}. If this macro is not defined, all such
3859 arguments are padded down if @code{BYTES_BIG_ENDIAN} is true.
3862 @defmac BLOCK_REG_PADDING (@var{mode}, @var{type}, @var{first})
3863 Specify padding for the last element of a block move between registers and
3864 memory. @var{first} is nonzero if this is the only element. Defining this
3865 macro allows better control of register function parameters on big-endian
3866 machines, without using @code{PARALLEL} rtl. In particular,
3867 @code{MUST_PASS_IN_STACK} need not test padding and mode of types in
3868 registers, as there is no longer a "wrong" part of a register; For example,
3869 a three byte aggregate may be passed in the high part of a register if so
3873 @defmac FUNCTION_ARG_BOUNDARY (@var{mode}, @var{type})
3874 If defined, a C expression that gives the alignment boundary, in bits,
3875 of an argument with the specified mode and type. If it is not defined,
3876 @code{PARM_BOUNDARY} is used for all arguments.
3879 @defmac FUNCTION_ARG_REGNO_P (@var{regno})
3880 A C expression that is nonzero if @var{regno} is the number of a hard
3881 register in which function arguments are sometimes passed. This does
3882 @emph{not} include implicit arguments such as the static chain and
3883 the structure-value address. On many machines, no registers can be
3884 used for this purpose since all function arguments are pushed on the
3888 @deftypefn {Target Hook} bool TARGET_SPLIT_COMPLEX_ARG (tree @var{type})
3889 This hook should return true if parameter of type @var{type} are passed
3890 as two scalar parameters. By default, GCC will attempt to pack complex
3891 arguments into the target's word size. Some ABIs require complex arguments
3892 to be split and treated as their individual components. For example, on
3893 AIX64, complex floats should be passed in a pair of floating point
3894 registers, even though a complex float would fit in one 64-bit floating
3897 The default value of this hook is @code{NULL}, which is treated as always
3901 @deftypefn {Target Hook} tree TARGET_BUILD_BUILTIN_VA_LIST (void)
3902 This hook returns a type node for @code{va_list} for the target.
3903 The default version of the hook returns @code{void*}.
3906 @deftypefn {Target Hook} tree TARGET_GIMPLIFY_VA_ARG_EXPR (tree @var{valist}, tree @var{type}, tree *@var{pre_p}, tree *@var{post_p})
3907 This hook performs target-specific gimplification of
3908 @code{VA_ARG_EXPR}. The first two parameters correspond to the
3909 arguments to @code{va_arg}; the latter two are as in
3910 @code{gimplify.c:gimplify_expr}.
3913 @deftypefn {Target Hook} bool TARGET_VALID_POINTER_MODE (enum machine_mode @var{mode})
3914 Define this to return nonzero if the port can handle pointers
3915 with machine mode @var{mode}. The default version of this
3916 hook returns true for both @code{ptr_mode} and @code{Pmode}.
3919 @deftypefn {Target Hook} bool TARGET_SCALAR_MODE_SUPPORTED_P (enum machine_mode @var{mode})
3920 Define this to return nonzero if the port is prepared to handle
3921 insns involving scalar mode @var{mode}. For a scalar mode to be
3922 considered supported, all the basic arithmetic and comparisons
3925 The default version of this hook returns true for any mode
3926 required to handle the basic C types (as defined by the port).
3927 Included here are the double-word arithmetic supported by the
3928 code in @file{optabs.c}.
3931 @deftypefn {Target Hook} bool TARGET_VECTOR_MODE_SUPPORTED_P (enum machine_mode @var{mode})
3932 Define this to return nonzero if the port is prepared to handle
3933 insns involving vector mode @var{mode}. At the very least, it
3934 must have move patterns for this mode.
3938 @subsection How Scalar Function Values Are Returned
3939 @cindex return values in registers
3940 @cindex values, returned by functions
3941 @cindex scalars, returned as values
3943 This section discusses the macros that control returning scalars as
3944 values---values that can fit in registers.
3946 @defmac FUNCTION_VALUE (@var{valtype}, @var{func})
3947 A C expression to create an RTX representing the place where a
3948 function returns a value of data type @var{valtype}. @var{valtype} is
3949 a tree node representing a data type. Write @code{TYPE_MODE
3950 (@var{valtype})} to get the machine mode used to represent that type.
3951 On many machines, only the mode is relevant. (Actually, on most
3952 machines, scalar values are returned in the same place regardless of
3955 The value of the expression is usually a @code{reg} RTX for the hard
3956 register where the return value is stored. The value can also be a
3957 @code{parallel} RTX, if the return value is in multiple places. See
3958 @code{FUNCTION_ARG} for an explanation of the @code{parallel} form.
3960 If @code{TARGET_PROMOTE_FUNCTION_RETURN} returns true, you must apply the same
3961 promotion rules specified in @code{PROMOTE_MODE} if @var{valtype} is a
3964 If the precise function being called is known, @var{func} is a tree
3965 node (@code{FUNCTION_DECL}) for it; otherwise, @var{func} is a null
3966 pointer. This makes it possible to use a different value-returning
3967 convention for specific functions when all their calls are
3970 @code{FUNCTION_VALUE} is not used for return vales with aggregate data
3971 types, because these are returned in another way. See
3972 @code{TARGET_STRUCT_VALUE_RTX} and related macros, below.
3975 @defmac FUNCTION_OUTGOING_VALUE (@var{valtype}, @var{func})
3976 Define this macro if the target machine has ``register windows''
3977 so that the register in which a function returns its value is not
3978 the same as the one in which the caller sees the value.
3980 For such machines, @code{FUNCTION_VALUE} computes the register in which
3981 the caller will see the value. @code{FUNCTION_OUTGOING_VALUE} should be
3982 defined in a similar fashion to tell the function where to put the
3985 If @code{FUNCTION_OUTGOING_VALUE} is not defined,
3986 @code{FUNCTION_VALUE} serves both purposes.
3988 @code{FUNCTION_OUTGOING_VALUE} is not used for return vales with
3989 aggregate data types, because these are returned in another way. See
3990 @code{TARGET_STRUCT_VALUE_RTX} and related macros, below.
3993 @defmac LIBCALL_VALUE (@var{mode})
3994 A C expression to create an RTX representing the place where a library
3995 function returns a value of mode @var{mode}. If the precise function
3996 being called is known, @var{func} is a tree node
3997 (@code{FUNCTION_DECL}) for it; otherwise, @var{func} is a null
3998 pointer. This makes it possible to use a different value-returning
3999 convention for specific functions when all their calls are
4002 Note that ``library function'' in this context means a compiler
4003 support routine, used to perform arithmetic, whose name is known
4004 specially by the compiler and was not mentioned in the C code being
4007 The definition of @code{LIBRARY_VALUE} need not be concerned aggregate
4008 data types, because none of the library functions returns such types.
4011 @defmac FUNCTION_VALUE_REGNO_P (@var{regno})
4012 A C expression that is nonzero if @var{regno} is the number of a hard
4013 register in which the values of called function may come back.
4015 A register whose use for returning values is limited to serving as the
4016 second of a pair (for a value of type @code{double}, say) need not be
4017 recognized by this macro. So for most machines, this definition
4021 #define FUNCTION_VALUE_REGNO_P(N) ((N) == 0)
4024 If the machine has register windows, so that the caller and the called
4025 function use different registers for the return value, this macro
4026 should recognize only the caller's register numbers.
4029 @defmac APPLY_RESULT_SIZE
4030 Define this macro if @samp{untyped_call} and @samp{untyped_return}
4031 need more space than is implied by @code{FUNCTION_VALUE_REGNO_P} for
4032 saving and restoring an arbitrary return value.
4035 @deftypefn {Target Hook} bool TARGET_RETURN_IN_MSB (tree @var{type})
4036 This hook should return true if values of type @var{type} are returned
4037 at the most significant end of a register (in other words, if they are
4038 padded at the least significant end). You can assume that @var{type}
4039 is returned in a register; the caller is required to check this.
4041 Note that the register provided by @code{FUNCTION_VALUE} must be able
4042 to hold the complete return value. For example, if a 1-, 2- or 3-byte
4043 structure is returned at the most significant end of a 4-byte register,
4044 @code{FUNCTION_VALUE} should provide an @code{SImode} rtx.
4047 @node Aggregate Return
4048 @subsection How Large Values Are Returned
4049 @cindex aggregates as return values
4050 @cindex large return values
4051 @cindex returning aggregate values
4052 @cindex structure value address
4054 When a function value's mode is @code{BLKmode} (and in some other
4055 cases), the value is not returned according to @code{FUNCTION_VALUE}
4056 (@pxref{Scalar Return}). Instead, the caller passes the address of a
4057 block of memory in which the value should be stored. This address
4058 is called the @dfn{structure value address}.
4060 This section describes how to control returning structure values in
4063 @deftypefn {Target Hook} bool TARGET_RETURN_IN_MEMORY (tree @var{type}, tree @var{fntype})
4064 This target hook should return a nonzero value to say to return the
4065 function value in memory, just as large structures are always returned.
4066 Here @var{type} will be the data type of the value, and @var{fntype}
4067 will be the type of the function doing the returning, or @code{NULL} for
4070 Note that values of mode @code{BLKmode} must be explicitly handled
4071 by this function. Also, the option @option{-fpcc-struct-return}
4072 takes effect regardless of this macro. On most systems, it is
4073 possible to leave the hook undefined; this causes a default
4074 definition to be used, whose value is the constant 1 for @code{BLKmode}
4075 values, and 0 otherwise.
4077 Do not use this hook to indicate that structures and unions should always
4078 be returned in memory. You should instead use @code{DEFAULT_PCC_STRUCT_RETURN}
4082 @defmac DEFAULT_PCC_STRUCT_RETURN
4083 Define this macro to be 1 if all structure and union return values must be
4084 in memory. Since this results in slower code, this should be defined
4085 only if needed for compatibility with other compilers or with an ABI@.
4086 If you define this macro to be 0, then the conventions used for structure
4087 and union return values are decided by the @code{TARGET_RETURN_IN_MEMORY}
4090 If not defined, this defaults to the value 1.
4093 @deftypefn {Target Hook} rtx TARGET_STRUCT_VALUE_RTX (tree @var{fndecl}, int @var{incoming})
4094 This target hook should return the location of the structure value
4095 address (normally a @code{mem} or @code{reg}), or 0 if the address is
4096 passed as an ``invisible'' first argument. Note that @var{fndecl} may
4097 be @code{NULL}, for libcalls. You do not need to define this target
4098 hook if the address is always passed as an ``invisible'' first
4101 On some architectures the place where the structure value address
4102 is found by the called function is not the same place that the
4103 caller put it. This can be due to register windows, or it could
4104 be because the function prologue moves it to a different place.
4105 @var{incoming} is @code{true} when the location is needed in
4106 the context of the called function, and @code{false} in the context of
4109 If @var{incoming} is @code{true} and the address is to be found on the
4110 stack, return a @code{mem} which refers to the frame pointer.
4113 @defmac PCC_STATIC_STRUCT_RETURN
4114 Define this macro if the usual system convention on the target machine
4115 for returning structures and unions is for the called function to return
4116 the address of a static variable containing the value.
4118 Do not define this if the usual system convention is for the caller to
4119 pass an address to the subroutine.
4121 This macro has effect in @option{-fpcc-struct-return} mode, but it does
4122 nothing when you use @option{-freg-struct-return} mode.
4126 @subsection Caller-Saves Register Allocation
4128 If you enable it, GCC can save registers around function calls. This
4129 makes it possible to use call-clobbered registers to hold variables that
4130 must live across calls.
4132 @defmac CALLER_SAVE_PROFITABLE (@var{refs}, @var{calls})
4133 A C expression to determine whether it is worthwhile to consider placing
4134 a pseudo-register in a call-clobbered hard register and saving and
4135 restoring it around each function call. The expression should be 1 when
4136 this is worth doing, and 0 otherwise.
4138 If you don't define this macro, a default is used which is good on most
4139 machines: @code{4 * @var{calls} < @var{refs}}.
4142 @defmac HARD_REGNO_CALLER_SAVE_MODE (@var{regno}, @var{nregs})
4143 A C expression specifying which mode is required for saving @var{nregs}
4144 of a pseudo-register in call-clobbered hard register @var{regno}. If
4145 @var{regno} is unsuitable for caller save, @code{VOIDmode} should be
4146 returned. For most machines this macro need not be defined since GCC
4147 will select the smallest suitable mode.
4150 @node Function Entry
4151 @subsection Function Entry and Exit
4152 @cindex function entry and exit
4156 This section describes the macros that output function entry
4157 (@dfn{prologue}) and exit (@dfn{epilogue}) code.
4159 @deftypefn {Target Hook} void TARGET_ASM_FUNCTION_PROLOGUE (FILE *@var{file}, HOST_WIDE_INT @var{size})
4160 If defined, a function that outputs the assembler code for entry to a
4161 function. The prologue is responsible for setting up the stack frame,
4162 initializing the frame pointer register, saving registers that must be
4163 saved, and allocating @var{size} additional bytes of storage for the
4164 local variables. @var{size} is an integer. @var{file} is a stdio
4165 stream to which the assembler code should be output.
4167 The label for the beginning of the function need not be output by this
4168 macro. That has already been done when the macro is run.
4170 @findex regs_ever_live
4171 To determine which registers to save, the macro can refer to the array
4172 @code{regs_ever_live}: element @var{r} is nonzero if hard register
4173 @var{r} is used anywhere within the function. This implies the function
4174 prologue should save register @var{r}, provided it is not one of the
4175 call-used registers. (@code{TARGET_ASM_FUNCTION_EPILOGUE} must likewise use
4176 @code{regs_ever_live}.)
4178 On machines that have ``register windows'', the function entry code does
4179 not save on the stack the registers that are in the windows, even if
4180 they are supposed to be preserved by function calls; instead it takes
4181 appropriate steps to ``push'' the register stack, if any non-call-used
4182 registers are used in the function.
4184 @findex frame_pointer_needed
4185 On machines where functions may or may not have frame-pointers, the
4186 function entry code must vary accordingly; it must set up the frame
4187 pointer if one is wanted, and not otherwise. To determine whether a
4188 frame pointer is in wanted, the macro can refer to the variable
4189 @code{frame_pointer_needed}. The variable's value will be 1 at run
4190 time in a function that needs a frame pointer. @xref{Elimination}.
4192 The function entry code is responsible for allocating any stack space
4193 required for the function. This stack space consists of the regions
4194 listed below. In most cases, these regions are allocated in the
4195 order listed, with the last listed region closest to the top of the
4196 stack (the lowest address if @code{STACK_GROWS_DOWNWARD} is defined, and
4197 the highest address if it is not defined). You can use a different order
4198 for a machine if doing so is more convenient or required for
4199 compatibility reasons. Except in cases where required by standard
4200 or by a debugger, there is no reason why the stack layout used by GCC
4201 need agree with that used by other compilers for a machine.
4204 @deftypefn {Target Hook} void TARGET_ASM_FUNCTION_END_PROLOGUE (FILE *@var{file})
4205 If defined, a function that outputs assembler code at the end of a
4206 prologue. This should be used when the function prologue is being
4207 emitted as RTL, and you have some extra assembler that needs to be
4208 emitted. @xref{prologue instruction pattern}.
4211 @deftypefn {Target Hook} void TARGET_ASM_FUNCTION_BEGIN_EPILOGUE (FILE *@var{file})
4212 If defined, a function that outputs assembler code at the start of an
4213 epilogue. This should be used when the function epilogue is being
4214 emitted as RTL, and you have some extra assembler that needs to be
4215 emitted. @xref{epilogue instruction pattern}.
4218 @deftypefn {Target Hook} void TARGET_ASM_FUNCTION_EPILOGUE (FILE *@var{file}, HOST_WIDE_INT @var{size})
4219 If defined, a function that outputs the assembler code for exit from a
4220 function. The epilogue is responsible for restoring the saved
4221 registers and stack pointer to their values when the function was
4222 called, and returning control to the caller. This macro takes the
4223 same arguments as the macro @code{TARGET_ASM_FUNCTION_PROLOGUE}, and the
4224 registers to restore are determined from @code{regs_ever_live} and
4225 @code{CALL_USED_REGISTERS} in the same way.
4227 On some machines, there is a single instruction that does all the work
4228 of returning from the function. On these machines, give that
4229 instruction the name @samp{return} and do not define the macro
4230 @code{TARGET_ASM_FUNCTION_EPILOGUE} at all.
4232 Do not define a pattern named @samp{return} if you want the
4233 @code{TARGET_ASM_FUNCTION_EPILOGUE} to be used. If you want the target
4234 switches to control whether return instructions or epilogues are used,
4235 define a @samp{return} pattern with a validity condition that tests the
4236 target switches appropriately. If the @samp{return} pattern's validity
4237 condition is false, epilogues will be used.
4239 On machines where functions may or may not have frame-pointers, the
4240 function exit code must vary accordingly. Sometimes the code for these
4241 two cases is completely different. To determine whether a frame pointer
4242 is wanted, the macro can refer to the variable
4243 @code{frame_pointer_needed}. The variable's value will be 1 when compiling
4244 a function that needs a frame pointer.
4246 Normally, @code{TARGET_ASM_FUNCTION_PROLOGUE} and
4247 @code{TARGET_ASM_FUNCTION_EPILOGUE} must treat leaf functions specially.
4248 The C variable @code{current_function_is_leaf} is nonzero for such a
4249 function. @xref{Leaf Functions}.
4251 On some machines, some functions pop their arguments on exit while
4252 others leave that for the caller to do. For example, the 68020 when
4253 given @option{-mrtd} pops arguments in functions that take a fixed
4254 number of arguments.
4256 @findex current_function_pops_args
4257 Your definition of the macro @code{RETURN_POPS_ARGS} decides which
4258 functions pop their own arguments. @code{TARGET_ASM_FUNCTION_EPILOGUE}
4259 needs to know what was decided. The variable that is called
4260 @code{current_function_pops_args} is the number of bytes of its
4261 arguments that a function should pop. @xref{Scalar Return}.
4262 @c what is the "its arguments" in the above sentence referring to, pray
4263 @c tell? --mew 5feb93
4268 @findex current_function_pretend_args_size
4269 A region of @code{current_function_pretend_args_size} bytes of
4270 uninitialized space just underneath the first argument arriving on the
4271 stack. (This may not be at the very start of the allocated stack region
4272 if the calling sequence has pushed anything else since pushing the stack
4273 arguments. But usually, on such machines, nothing else has been pushed
4274 yet, because the function prologue itself does all the pushing.) This
4275 region is used on machines where an argument may be passed partly in
4276 registers and partly in memory, and, in some cases to support the
4277 features in @code{<stdarg.h>}.
4280 An area of memory used to save certain registers used by the function.
4281 The size of this area, which may also include space for such things as
4282 the return address and pointers to previous stack frames, is
4283 machine-specific and usually depends on which registers have been used
4284 in the function. Machines with register windows often do not require
4288 A region of at least @var{size} bytes, possibly rounded up to an allocation
4289 boundary, to contain the local variables of the function. On some machines,
4290 this region and the save area may occur in the opposite order, with the
4291 save area closer to the top of the stack.
4294 @cindex @code{ACCUMULATE_OUTGOING_ARGS} and stack frames
4295 Optionally, when @code{ACCUMULATE_OUTGOING_ARGS} is defined, a region of
4296 @code{current_function_outgoing_args_size} bytes to be used for outgoing
4297 argument lists of the function. @xref{Stack Arguments}.
4300 @defmac EXIT_IGNORE_STACK
4301 Define this macro as a C expression that is nonzero if the return
4302 instruction or the function epilogue ignores the value of the stack
4303 pointer; in other words, if it is safe to delete an instruction to
4304 adjust the stack pointer before a return from the function. The
4307 Note that this macro's value is relevant only for functions for which
4308 frame pointers are maintained. It is never safe to delete a final
4309 stack adjustment in a function that has no frame pointer, and the
4310 compiler knows this regardless of @code{EXIT_IGNORE_STACK}.
4313 @defmac EPILOGUE_USES (@var{regno})
4314 Define this macro as a C expression that is nonzero for registers that are
4315 used by the epilogue or the @samp{return} pattern. The stack and frame
4316 pointer registers are already be assumed to be used as needed.
4319 @defmac EH_USES (@var{regno})
4320 Define this macro as a C expression that is nonzero for registers that are
4321 used by the exception handling mechanism, and so should be considered live
4322 on entry to an exception edge.
4325 @defmac DELAY_SLOTS_FOR_EPILOGUE
4326 Define this macro if the function epilogue contains delay slots to which
4327 instructions from the rest of the function can be ``moved''. The
4328 definition should be a C expression whose value is an integer
4329 representing the number of delay slots there.
4332 @defmac ELIGIBLE_FOR_EPILOGUE_DELAY (@var{insn}, @var{n})
4333 A C expression that returns 1 if @var{insn} can be placed in delay
4334 slot number @var{n} of the epilogue.
4336 The argument @var{n} is an integer which identifies the delay slot now
4337 being considered (since different slots may have different rules of
4338 eligibility). It is never negative and is always less than the number
4339 of epilogue delay slots (what @code{DELAY_SLOTS_FOR_EPILOGUE} returns).
4340 If you reject a particular insn for a given delay slot, in principle, it
4341 may be reconsidered for a subsequent delay slot. Also, other insns may
4342 (at least in principle) be considered for the so far unfilled delay
4345 @findex current_function_epilogue_delay_list
4346 @findex final_scan_insn
4347 The insns accepted to fill the epilogue delay slots are put in an RTL
4348 list made with @code{insn_list} objects, stored in the variable
4349 @code{current_function_epilogue_delay_list}. The insn for the first
4350 delay slot comes first in the list. Your definition of the macro
4351 @code{TARGET_ASM_FUNCTION_EPILOGUE} should fill the delay slots by
4352 outputting the insns in this list, usually by calling
4353 @code{final_scan_insn}.
4355 You need not define this macro if you did not define
4356 @code{DELAY_SLOTS_FOR_EPILOGUE}.
4359 @deftypefn {Target Hook} void TARGET_ASM_OUTPUT_MI_THUNK (FILE *@var{file}, tree @var{thunk_fndecl}, HOST_WIDE_INT @var{delta}, HOST_WIDE_INT @var{vcall_offset}, tree @var{function})
4360 A function that outputs the assembler code for a thunk
4361 function, used to implement C++ virtual function calls with multiple
4362 inheritance. The thunk acts as a wrapper around a virtual function,
4363 adjusting the implicit object parameter before handing control off to
4366 First, emit code to add the integer @var{delta} to the location that
4367 contains the incoming first argument. Assume that this argument
4368 contains a pointer, and is the one used to pass the @code{this} pointer
4369 in C++. This is the incoming argument @emph{before} the function prologue,
4370 e.g.@: @samp{%o0} on a sparc. The addition must preserve the values of
4371 all other incoming arguments.
4373 Then, if @var{vcall_offset} is nonzero, an additional adjustment should be
4374 made after adding @code{delta}. In particular, if @var{p} is the
4375 adjusted pointer, the following adjustment should be made:
4378 p += (*((ptrdiff_t **)p))[vcall_offset/sizeof(ptrdiff_t)]
4381 After the additions, emit code to jump to @var{function}, which is a
4382 @code{FUNCTION_DECL}. This is a direct pure jump, not a call, and does
4383 not touch the return address. Hence returning from @var{FUNCTION} will
4384 return to whoever called the current @samp{thunk}.
4386 The effect must be as if @var{function} had been called directly with
4387 the adjusted first argument. This macro is responsible for emitting all
4388 of the code for a thunk function; @code{TARGET_ASM_FUNCTION_PROLOGUE}
4389 and @code{TARGET_ASM_FUNCTION_EPILOGUE} are not invoked.
4391 The @var{thunk_fndecl} is redundant. (@var{delta} and @var{function}
4392 have already been extracted from it.) It might possibly be useful on
4393 some targets, but probably not.
4395 If you do not define this macro, the target-independent code in the C++
4396 front end will generate a less efficient heavyweight thunk that calls
4397 @var{function} instead of jumping to it. The generic approach does
4398 not support varargs.
4401 @deftypefn {Target Hook} bool TARGET_ASM_CAN_OUTPUT_MI_THUNK (tree @var{thunk_fndecl}, HOST_WIDE_INT @var{delta}, HOST_WIDE_INT @var{vcall_offset}, tree @var{function})
4402 A function that returns true if TARGET_ASM_OUTPUT_MI_THUNK would be able
4403 to output the assembler code for the thunk function specified by the
4404 arguments it is passed, and false otherwise. In the latter case, the
4405 generic approach will be used by the C++ front end, with the limitations
4410 @subsection Generating Code for Profiling
4411 @cindex profiling, code generation
4413 These macros will help you generate code for profiling.
4415 @defmac FUNCTION_PROFILER (@var{file}, @var{labelno})
4416 A C statement or compound statement to output to @var{file} some
4417 assembler code to call the profiling subroutine @code{mcount}.
4420 The details of how @code{mcount} expects to be called are determined by
4421 your operating system environment, not by GCC@. To figure them out,
4422 compile a small program for profiling using the system's installed C
4423 compiler and look at the assembler code that results.
4425 Older implementations of @code{mcount} expect the address of a counter
4426 variable to be loaded into some register. The name of this variable is
4427 @samp{LP} followed by the number @var{labelno}, so you would generate
4428 the name using @samp{LP%d} in a @code{fprintf}.
4431 @defmac PROFILE_HOOK
4432 A C statement or compound statement to output to @var{file} some assembly
4433 code to call the profiling subroutine @code{mcount} even the target does
4434 not support profiling.
4437 @defmac NO_PROFILE_COUNTERS
4438 Define this macro if the @code{mcount} subroutine on your system does
4439 not need a counter variable allocated for each function. This is true
4440 for almost all modern implementations. If you define this macro, you
4441 must not use the @var{labelno} argument to @code{FUNCTION_PROFILER}.
4444 @defmac PROFILE_BEFORE_PROLOGUE
4445 Define this macro if the code for function profiling should come before
4446 the function prologue. Normally, the profiling code comes after.
4450 @subsection Permitting tail calls
4453 @deftypefn {Target Hook} bool TARGET_FUNCTION_OK_FOR_SIBCALL (tree @var{decl}, tree @var{exp})
4454 True if it is ok to do sibling call optimization for the specified
4455 call expression @var{exp}. @var{decl} will be the called function,
4456 or @code{NULL} if this is an indirect call.
4458 It is not uncommon for limitations of calling conventions to prevent
4459 tail calls to functions outside the current unit of translation, or
4460 during PIC compilation. The hook is used to enforce these restrictions,
4461 as the @code{sibcall} md pattern can not fail, or fall over to a
4462 ``normal'' call. The criteria for successful sibling call optimization
4463 may vary greatly between different architectures.
4466 @node Stack Smashing Protection
4467 @subsection Stack smashing protection
4468 @cindex stack smashing protection
4470 @deftypefn {Target Hook} tree TARGET_STACK_PROTECT_GUARD (void)
4471 This hook returns a @code{DECL} node for the external variable to use
4472 for the stack protection guard. This variable is initialized by the
4473 runtime to some random value and is used to initialize the guard value
4474 that is placed at the top of the local stack frame. The type of this
4475 variable must be @code{ptr_type_node}.
4477 The default version of this hook creates a variable called
4478 @samp{__stack_chk_guard}, which is normally defined in @file{libgcc2.c}.
4481 @deftypefn {Target Hook} tree TARGET_STACK_PROTECT_FAIL (void)
4482 This hook returns a tree expression that alerts the runtime that the
4483 stack protect guard variable has been modified. This expression should
4484 involve a call to a @code{noreturn} function.
4486 The default version of this hook invokes a function called
4487 @samp{__stack_chk_fail}, taking no arguments. This function is
4488 normally defined in @file{libgcc2.c}.
4492 @section Implementing the Varargs Macros
4493 @cindex varargs implementation
4495 GCC comes with an implementation of @code{<varargs.h>} and
4496 @code{<stdarg.h>} that work without change on machines that pass arguments
4497 on the stack. Other machines require their own implementations of
4498 varargs, and the two machine independent header files must have
4499 conditionals to include it.
4501 ISO @code{<stdarg.h>} differs from traditional @code{<varargs.h>} mainly in
4502 the calling convention for @code{va_start}. The traditional
4503 implementation takes just one argument, which is the variable in which
4504 to store the argument pointer. The ISO implementation of
4505 @code{va_start} takes an additional second argument. The user is
4506 supposed to write the last named argument of the function here.
4508 However, @code{va_start} should not use this argument. The way to find
4509 the end of the named arguments is with the built-in functions described
4512 @defmac __builtin_saveregs ()
4513 Use this built-in function to save the argument registers in memory so
4514 that the varargs mechanism can access them. Both ISO and traditional
4515 versions of @code{va_start} must use @code{__builtin_saveregs}, unless
4516 you use @code{TARGET_SETUP_INCOMING_VARARGS} (see below) instead.
4518 On some machines, @code{__builtin_saveregs} is open-coded under the
4519 control of the target hook @code{TARGET_EXPAND_BUILTIN_SAVEREGS}. On
4520 other machines, it calls a routine written in assembler language,
4521 found in @file{libgcc2.c}.
4523 Code generated for the call to @code{__builtin_saveregs} appears at the
4524 beginning of the function, as opposed to where the call to
4525 @code{__builtin_saveregs} is written, regardless of what the code is.
4526 This is because the registers must be saved before the function starts
4527 to use them for its own purposes.
4528 @c i rewrote the first sentence above to fix an overfull hbox. --mew
4532 @defmac __builtin_args_info (@var{category})
4533 Use this built-in function to find the first anonymous arguments in
4536 In general, a machine may have several categories of registers used for
4537 arguments, each for a particular category of data types. (For example,
4538 on some machines, floating-point registers are used for floating-point
4539 arguments while other arguments are passed in the general registers.)
4540 To make non-varargs functions use the proper calling convention, you
4541 have defined the @code{CUMULATIVE_ARGS} data type to record how many
4542 registers in each category have been used so far
4544 @code{__builtin_args_info} accesses the same data structure of type
4545 @code{CUMULATIVE_ARGS} after the ordinary argument layout is finished
4546 with it, with @var{category} specifying which word to access. Thus, the
4547 value indicates the first unused register in a given category.
4549 Normally, you would use @code{__builtin_args_info} in the implementation
4550 of @code{va_start}, accessing each category just once and storing the
4551 value in the @code{va_list} object. This is because @code{va_list} will
4552 have to update the values, and there is no way to alter the
4553 values accessed by @code{__builtin_args_info}.
4556 @defmac __builtin_next_arg (@var{lastarg})
4557 This is the equivalent of @code{__builtin_args_info}, for stack
4558 arguments. It returns the address of the first anonymous stack
4559 argument, as type @code{void *}. If @code{ARGS_GROW_DOWNWARD}, it
4560 returns the address of the location above the first anonymous stack
4561 argument. Use it in @code{va_start} to initialize the pointer for
4562 fetching arguments from the stack. Also use it in @code{va_start} to
4563 verify that the second parameter @var{lastarg} is the last named argument
4564 of the current function.
4567 @defmac __builtin_classify_type (@var{object})
4568 Since each machine has its own conventions for which data types are
4569 passed in which kind of register, your implementation of @code{va_arg}
4570 has to embody these conventions. The easiest way to categorize the
4571 specified data type is to use @code{__builtin_classify_type} together
4572 with @code{sizeof} and @code{__alignof__}.
4574 @code{__builtin_classify_type} ignores the value of @var{object},
4575 considering only its data type. It returns an integer describing what
4576 kind of type that is---integer, floating, pointer, structure, and so on.
4578 The file @file{typeclass.h} defines an enumeration that you can use to
4579 interpret the values of @code{__builtin_classify_type}.
4582 These machine description macros help implement varargs:
4584 @deftypefn {Target Hook} rtx TARGET_EXPAND_BUILTIN_SAVEREGS (void)
4585 If defined, this hook produces the machine-specific code for a call to
4586 @code{__builtin_saveregs}. This code will be moved to the very
4587 beginning of the function, before any parameter access are made. The
4588 return value of this function should be an RTX that contains the value
4589 to use as the return of @code{__builtin_saveregs}.
4592 @deftypefn {Target Hook} void TARGET_SETUP_INCOMING_VARARGS (CUMULATIVE_ARGS *@var{args_so_far}, enum machine_mode @var{mode}, tree @var{type}, int *@var{pretend_args_size}, int @var{second_time})
4593 This target hook offers an alternative to using
4594 @code{__builtin_saveregs} and defining the hook
4595 @code{TARGET_EXPAND_BUILTIN_SAVEREGS}. Use it to store the anonymous
4596 register arguments into the stack so that all the arguments appear to
4597 have been passed consecutively on the stack. Once this is done, you can
4598 use the standard implementation of varargs that works for machines that
4599 pass all their arguments on the stack.
4601 The argument @var{args_so_far} points to the @code{CUMULATIVE_ARGS} data
4602 structure, containing the values that are obtained after processing the
4603 named arguments. The arguments @var{mode} and @var{type} describe the
4604 last named argument---its machine mode and its data type as a tree node.
4606 The target hook should do two things: first, push onto the stack all the
4607 argument registers @emph{not} used for the named arguments, and second,
4608 store the size of the data thus pushed into the @code{int}-valued
4609 variable pointed to by @var{pretend_args_size}. The value that you
4610 store here will serve as additional offset for setting up the stack
4613 Because you must generate code to push the anonymous arguments at
4614 compile time without knowing their data types,
4615 @code{TARGET_SETUP_INCOMING_VARARGS} is only useful on machines that
4616 have just a single category of argument register and use it uniformly
4619 If the argument @var{second_time} is nonzero, it means that the
4620 arguments of the function are being analyzed for the second time. This
4621 happens for an inline function, which is not actually compiled until the
4622 end of the source file. The hook @code{TARGET_SETUP_INCOMING_VARARGS} should
4623 not generate any instructions in this case.
4626 @deftypefn {Target Hook} bool TARGET_STRICT_ARGUMENT_NAMING (CUMULATIVE_ARGS *@var{ca})
4627 Define this hook to return @code{true} if the location where a function
4628 argument is passed depends on whether or not it is a named argument.
4630 This hook controls how the @var{named} argument to @code{FUNCTION_ARG}
4631 is set for varargs and stdarg functions. If this hook returns
4632 @code{true}, the @var{named} argument is always true for named
4633 arguments, and false for unnamed arguments. If it returns @code{false},
4634 but @code{TARGET_PRETEND_OUTGOING_VARARGS_NAMED} returns @code{true},
4635 then all arguments are treated as named. Otherwise, all named arguments
4636 except the last are treated as named.
4638 You need not define this hook if it always returns zero.
4641 @deftypefn {Target Hook} bool TARGET_PRETEND_OUTGOING_VARARGS_NAMED
4642 If you need to conditionally change ABIs so that one works with
4643 @code{TARGET_SETUP_INCOMING_VARARGS}, but the other works like neither
4644 @code{TARGET_SETUP_INCOMING_VARARGS} nor @code{TARGET_STRICT_ARGUMENT_NAMING} was
4645 defined, then define this hook to return @code{true} if
4646 @code{TARGET_SETUP_INCOMING_VARARGS} is used, @code{false} otherwise.
4647 Otherwise, you should not define this hook.
4651 @section Trampolines for Nested Functions
4652 @cindex trampolines for nested functions
4653 @cindex nested functions, trampolines for
4655 A @dfn{trampoline} is a small piece of code that is created at run time
4656 when the address of a nested function is taken. It normally resides on
4657 the stack, in the stack frame of the containing function. These macros
4658 tell GCC how to generate code to allocate and initialize a
4661 The instructions in the trampoline must do two things: load a constant
4662 address into the static chain register, and jump to the real address of
4663 the nested function. On CISC machines such as the m68k, this requires
4664 two instructions, a move immediate and a jump. Then the two addresses
4665 exist in the trampoline as word-long immediate operands. On RISC
4666 machines, it is often necessary to load each address into a register in
4667 two parts. Then pieces of each address form separate immediate
4670 The code generated to initialize the trampoline must store the variable
4671 parts---the static chain value and the function address---into the
4672 immediate operands of the instructions. On a CISC machine, this is
4673 simply a matter of copying each address to a memory reference at the
4674 proper offset from the start of the trampoline. On a RISC machine, it
4675 may be necessary to take out pieces of the address and store them
4678 @defmac TRAMPOLINE_TEMPLATE (@var{file})
4679 A C statement to output, on the stream @var{file}, assembler code for a
4680 block of data that contains the constant parts of a trampoline. This
4681 code should not include a label---the label is taken care of
4684 If you do not define this macro, it means no template is needed
4685 for the target. Do not define this macro on systems where the block move
4686 code to copy the trampoline into place would be larger than the code
4687 to generate it on the spot.
4690 @defmac TRAMPOLINE_SECTION
4691 The name of a subroutine to switch to the section in which the
4692 trampoline template is to be placed (@pxref{Sections}). The default is
4693 a value of @samp{readonly_data_section}, which places the trampoline in
4694 the section containing read-only data.
4697 @defmac TRAMPOLINE_SIZE
4698 A C expression for the size in bytes of the trampoline, as an integer.
4701 @defmac TRAMPOLINE_ALIGNMENT
4702 Alignment required for trampolines, in bits.
4704 If you don't define this macro, the value of @code{BIGGEST_ALIGNMENT}
4705 is used for aligning trampolines.
4708 @defmac INITIALIZE_TRAMPOLINE (@var{addr}, @var{fnaddr}, @var{static_chain})
4709 A C statement to initialize the variable parts of a trampoline.
4710 @var{addr} is an RTX for the address of the trampoline; @var{fnaddr} is
4711 an RTX for the address of the nested function; @var{static_chain} is an
4712 RTX for the static chain value that should be passed to the function
4716 @defmac TRAMPOLINE_ADJUST_ADDRESS (@var{addr})
4717 A C statement that should perform any machine-specific adjustment in
4718 the address of the trampoline. Its argument contains the address that
4719 was passed to @code{INITIALIZE_TRAMPOLINE}. In case the address to be
4720 used for a function call should be different from the address in which
4721 the template was stored, the different address should be assigned to
4722 @var{addr}. If this macro is not defined, @var{addr} will be used for
4725 @cindex @code{TARGET_ASM_FUNCTION_EPILOGUE} and trampolines
4726 @cindex @code{TARGET_ASM_FUNCTION_PROLOGUE} and trampolines
4727 If this macro is not defined, by default the trampoline is allocated as
4728 a stack slot. This default is right for most machines. The exceptions
4729 are machines where it is impossible to execute instructions in the stack
4730 area. On such machines, you may have to implement a separate stack,
4731 using this macro in conjunction with @code{TARGET_ASM_FUNCTION_PROLOGUE}
4732 and @code{TARGET_ASM_FUNCTION_EPILOGUE}.
4734 @var{fp} points to a data structure, a @code{struct function}, which
4735 describes the compilation status of the immediate containing function of
4736 the function which the trampoline is for. The stack slot for the
4737 trampoline is in the stack frame of this containing function. Other
4738 allocation strategies probably must do something analogous with this
4742 Implementing trampolines is difficult on many machines because they have
4743 separate instruction and data caches. Writing into a stack location
4744 fails to clear the memory in the instruction cache, so when the program
4745 jumps to that location, it executes the old contents.
4747 Here are two possible solutions. One is to clear the relevant parts of
4748 the instruction cache whenever a trampoline is set up. The other is to
4749 make all trampolines identical, by having them jump to a standard
4750 subroutine. The former technique makes trampoline execution faster; the
4751 latter makes initialization faster.
4753 To clear the instruction cache when a trampoline is initialized, define
4754 the following macro.
4756 @defmac CLEAR_INSN_CACHE (@var{beg}, @var{end})
4757 If defined, expands to a C expression clearing the @emph{instruction
4758 cache} in the specified interval. The definition of this macro would
4759 typically be a series of @code{asm} statements. Both @var{beg} and
4760 @var{end} are both pointer expressions.
4763 The operating system may also require the stack to be made executable
4764 before calling the trampoline. To implement this requirement, define
4765 the following macro.
4767 @defmac ENABLE_EXECUTE_STACK
4768 Define this macro if certain operations must be performed before executing
4769 code located on the stack. The macro should expand to a series of C
4770 file-scope constructs (e.g.@: functions) and provide a unique entry point
4771 named @code{__enable_execute_stack}. The target is responsible for
4772 emitting calls to the entry point in the code, for example from the
4773 @code{INITIALIZE_TRAMPOLINE} macro.
4776 To use a standard subroutine, define the following macro. In addition,
4777 you must make sure that the instructions in a trampoline fill an entire
4778 cache line with identical instructions, or else ensure that the
4779 beginning of the trampoline code is always aligned at the same point in
4780 its cache line. Look in @file{m68k.h} as a guide.
4782 @defmac TRANSFER_FROM_TRAMPOLINE
4783 Define this macro if trampolines need a special subroutine to do their
4784 work. The macro should expand to a series of @code{asm} statements
4785 which will be compiled with GCC@. They go in a library function named
4786 @code{__transfer_from_trampoline}.
4788 If you need to avoid executing the ordinary prologue code of a compiled
4789 C function when you jump to the subroutine, you can do so by placing a
4790 special label of your own in the assembler code. Use one @code{asm}
4791 statement to generate an assembler label, and another to make the label
4792 global. Then trampolines can use that label to jump directly to your
4793 special assembler code.
4797 @section Implicit Calls to Library Routines
4798 @cindex library subroutine names
4799 @cindex @file{libgcc.a}
4801 @c prevent bad page break with this line
4802 Here is an explanation of implicit calls to library routines.
4804 @defmac DECLARE_LIBRARY_RENAMES
4805 This macro, if defined, should expand to a piece of C code that will get
4806 expanded when compiling functions for libgcc.a. It can be used to
4807 provide alternate names for GCC's internal library functions if there
4808 are ABI-mandated names that the compiler should provide.
4811 @findex init_one_libfunc
4812 @findex set_optab_libfunc
4813 @deftypefn {Target Hook} void TARGET_INIT_LIBFUNCS (void)
4814 This hook should declare additional library routines or rename
4815 existing ones, using the functions @code{set_optab_libfunc} and
4816 @code{init_one_libfunc} defined in @file{optabs.c}.
4817 @code{init_optabs} calls this macro after initializing all the normal
4820 The default is to do nothing. Most ports don't need to define this hook.
4823 @defmac FLOAT_LIB_COMPARE_RETURNS_BOOL (@var{mode}, @var{comparison})
4824 This macro should return @code{true} if the library routine that
4825 implements the floating point comparison operator @var{comparison} in
4826 mode @var{mode} will return a boolean, and @var{false} if it will
4829 GCC's own floating point libraries return tristates from the
4830 comparison operators, so the default returns false always. Most ports
4831 don't need to define this macro.
4834 @defmac TARGET_LIB_INT_CMP_BIASED
4835 This macro should evaluate to @code{true} if the integer comparison
4836 functions (like @code{__cmpdi2}) return 0 to indicate that the first
4837 operand is smaller than the second, 1 to indicate that they are equal,
4838 and 2 to indicate that the first operand is greater than the second.
4839 If this macro evaluates to @code{false} the comparison functions return
4840 @minus{}1, 0, and 1 instead of 0, 1, and 2. If the target uses the routines
4841 in @file{libgcc.a}, you do not need to define this macro.
4844 @cindex US Software GOFAST, floating point emulation library
4845 @cindex floating point emulation library, US Software GOFAST
4846 @cindex GOFAST, floating point emulation library
4847 @findex gofast_maybe_init_libfuncs
4848 @defmac US_SOFTWARE_GOFAST
4849 Define this macro if your system C library uses the US Software GOFAST
4850 library to provide floating point emulation.
4852 In addition to defining this macro, your architecture must set
4853 @code{TARGET_INIT_LIBFUNCS} to @code{gofast_maybe_init_libfuncs}, or
4854 else call that function from its version of that hook. It is defined
4855 in @file{config/gofast.h}, which must be included by your
4856 architecture's @file{@var{cpu}.c} file. See @file{sparc/sparc.c} for
4859 If this macro is defined, the
4860 @code{TARGET_FLOAT_LIB_COMPARE_RETURNS_BOOL} target hook must return
4861 false for @code{SFmode} and @code{DFmode} comparisons.
4864 @cindex @code{EDOM}, implicit usage
4867 The value of @code{EDOM} on the target machine, as a C integer constant
4868 expression. If you don't define this macro, GCC does not attempt to
4869 deposit the value of @code{EDOM} into @code{errno} directly. Look in
4870 @file{/usr/include/errno.h} to find the value of @code{EDOM} on your
4873 If you do not define @code{TARGET_EDOM}, then compiled code reports
4874 domain errors by calling the library function and letting it report the
4875 error. If mathematical functions on your system use @code{matherr} when
4876 there is an error, then you should leave @code{TARGET_EDOM} undefined so
4877 that @code{matherr} is used normally.
4880 @cindex @code{errno}, implicit usage
4881 @defmac GEN_ERRNO_RTX
4882 Define this macro as a C expression to create an rtl expression that
4883 refers to the global ``variable'' @code{errno}. (On certain systems,
4884 @code{errno} may not actually be a variable.) If you don't define this
4885 macro, a reasonable default is used.
4888 @cindex C99 math functions, implicit usage
4889 @defmac TARGET_C99_FUNCTIONS
4890 When this macro is nonzero, GCC will implicitly optimize @code{sin} calls into
4891 @code{sinf} and similarly for other functions defined by C99 standard. The
4892 default is nonzero that should be proper value for most modern systems, however
4893 number of existing systems lacks support for these functions in the runtime so
4894 they needs this macro to be redefined to 0.
4897 @defmac NEXT_OBJC_RUNTIME
4898 Define this macro to generate code for Objective-C message sending using
4899 the calling convention of the NeXT system. This calling convention
4900 involves passing the object, the selector and the method arguments all
4901 at once to the method-lookup library function.
4903 The default calling convention passes just the object and the selector
4904 to the lookup function, which returns a pointer to the method.
4907 @node Addressing Modes
4908 @section Addressing Modes
4909 @cindex addressing modes
4911 @c prevent bad page break with this line
4912 This is about addressing modes.
4914 @defmac HAVE_PRE_INCREMENT
4915 @defmacx HAVE_PRE_DECREMENT
4916 @defmacx HAVE_POST_INCREMENT
4917 @defmacx HAVE_POST_DECREMENT
4918 A C expression that is nonzero if the machine supports pre-increment,
4919 pre-decrement, post-increment, or post-decrement addressing respectively.
4922 @defmac HAVE_PRE_MODIFY_DISP
4923 @defmacx HAVE_POST_MODIFY_DISP
4924 A C expression that is nonzero if the machine supports pre- or
4925 post-address side-effect generation involving constants other than
4926 the size of the memory operand.
4929 @defmac HAVE_PRE_MODIFY_REG
4930 @defmacx HAVE_POST_MODIFY_REG
4931 A C expression that is nonzero if the machine supports pre- or
4932 post-address side-effect generation involving a register displacement.
4935 @defmac CONSTANT_ADDRESS_P (@var{x})
4936 A C expression that is 1 if the RTX @var{x} is a constant which
4937 is a valid address. On most machines, this can be defined as
4938 @code{CONSTANT_P (@var{x})}, but a few machines are more restrictive
4939 in which constant addresses are supported.
4942 @defmac CONSTANT_P (@var{x})
4943 @code{CONSTANT_P}, which is defined by target-independent code,
4944 accepts integer-values expressions whose values are not explicitly
4945 known, such as @code{symbol_ref}, @code{label_ref}, and @code{high}
4946 expressions and @code{const} arithmetic expressions, in addition to
4947 @code{const_int} and @code{const_double} expressions.
4950 @defmac MAX_REGS_PER_ADDRESS
4951 A number, the maximum number of registers that can appear in a valid
4952 memory address. Note that it is up to you to specify a value equal to
4953 the maximum number that @code{GO_IF_LEGITIMATE_ADDRESS} would ever
4957 @defmac GO_IF_LEGITIMATE_ADDRESS (@var{mode}, @var{x}, @var{label})
4958 A C compound statement with a conditional @code{goto @var{label};}
4959 executed if @var{x} (an RTX) is a legitimate memory address on the
4960 target machine for a memory operand of mode @var{mode}.
4962 It usually pays to define several simpler macros to serve as
4963 subroutines for this one. Otherwise it may be too complicated to
4966 This macro must exist in two variants: a strict variant and a
4967 non-strict one. The strict variant is used in the reload pass. It
4968 must be defined so that any pseudo-register that has not been
4969 allocated a hard register is considered a memory reference. In
4970 contexts where some kind of register is required, a pseudo-register
4971 with no hard register must be rejected.
4973 The non-strict variant is used in other passes. It must be defined to
4974 accept all pseudo-registers in every context where some kind of
4975 register is required.
4977 @findex REG_OK_STRICT
4978 Compiler source files that want to use the strict variant of this
4979 macro define the macro @code{REG_OK_STRICT}. You should use an
4980 @code{#ifdef REG_OK_STRICT} conditional to define the strict variant
4981 in that case and the non-strict variant otherwise.
4983 Subroutines to check for acceptable registers for various purposes (one
4984 for base registers, one for index registers, and so on) are typically
4985 among the subroutines used to define @code{GO_IF_LEGITIMATE_ADDRESS}.
4986 Then only these subroutine macros need have two variants; the higher
4987 levels of macros may be the same whether strict or not.
4989 Normally, constant addresses which are the sum of a @code{symbol_ref}
4990 and an integer are stored inside a @code{const} RTX to mark them as
4991 constant. Therefore, there is no need to recognize such sums
4992 specifically as legitimate addresses. Normally you would simply
4993 recognize any @code{const} as legitimate.
4995 Usually @code{PRINT_OPERAND_ADDRESS} is not prepared to handle constant
4996 sums that are not marked with @code{const}. It assumes that a naked
4997 @code{plus} indicates indexing. If so, then you @emph{must} reject such
4998 naked constant sums as illegitimate addresses, so that none of them will
4999 be given to @code{PRINT_OPERAND_ADDRESS}.
5001 @cindex @code{TARGET_ENCODE_SECTION_INFO} and address validation
5002 On some machines, whether a symbolic address is legitimate depends on
5003 the section that the address refers to. On these machines, define the
5004 target hook @code{TARGET_ENCODE_SECTION_INFO} to store the information
5005 into the @code{symbol_ref}, and then check for it here. When you see a
5006 @code{const}, you will have to look inside it to find the
5007 @code{symbol_ref} in order to determine the section. @xref{Assembler
5011 @defmac REG_OK_FOR_BASE_P (@var{x})
5012 A C expression that is nonzero if @var{x} (assumed to be a @code{reg}
5013 RTX) is valid for use as a base register. For hard registers, it
5014 should always accept those which the hardware permits and reject the
5015 others. Whether the macro accepts or rejects pseudo registers must be
5016 controlled by @code{REG_OK_STRICT} as described above. This usually
5017 requires two variant definitions, of which @code{REG_OK_STRICT}
5018 controls the one actually used.
5021 @defmac REG_MODE_OK_FOR_BASE_P (@var{x}, @var{mode})
5022 A C expression that is just like @code{REG_OK_FOR_BASE_P}, except that
5023 that expression may examine the mode of the memory reference in
5024 @var{mode}. You should define this macro if the mode of the memory
5025 reference affects whether a register may be used as a base register. If
5026 you define this macro, the compiler will use it instead of
5027 @code{REG_OK_FOR_BASE_P}.
5030 @defmac REG_MODE_OK_FOR_REG_BASE_P (@var{x}, @var{mode})
5031 A C expression which is nonzero if @var{x} (assumed to be a @code{reg} RTX)
5032 is suitable for use as a base register in base plus index operand addresses,
5033 accessing memory in mode @var{mode}. It may be either a suitable hard
5034 register or a pseudo register that has been allocated such a hard register.
5035 You should define this macro if base plus index addresses have different
5036 requirements than other base register uses.
5039 @defmac REG_OK_FOR_INDEX_P (@var{x})
5040 A C expression that is nonzero if @var{x} (assumed to be a @code{reg}
5041 RTX) is valid for use as an index register.
5043 The difference between an index register and a base register is that
5044 the index register may be scaled. If an address involves the sum of
5045 two registers, neither one of them scaled, then either one may be
5046 labeled the ``base'' and the other the ``index''; but whichever
5047 labeling is used must fit the machine's constraints of which registers
5048 may serve in each capacity. The compiler will try both labelings,
5049 looking for one that is valid, and will reload one or both registers
5050 only if neither labeling works.
5053 @defmac FIND_BASE_TERM (@var{x})
5054 A C expression to determine the base term of address @var{x}.
5055 This macro is used in only one place: `find_base_term' in alias.c.
5057 It is always safe for this macro to not be defined. It exists so
5058 that alias analysis can understand machine-dependent addresses.
5060 The typical use of this macro is to handle addresses containing
5061 a label_ref or symbol_ref within an UNSPEC@.
5064 @defmac LEGITIMIZE_ADDRESS (@var{x}, @var{oldx}, @var{mode}, @var{win})
5065 A C compound statement that attempts to replace @var{x} with a valid
5066 memory address for an operand of mode @var{mode}. @var{win} will be a
5067 C statement label elsewhere in the code; the macro definition may use
5070 GO_IF_LEGITIMATE_ADDRESS (@var{mode}, @var{x}, @var{win});
5074 to avoid further processing if the address has become legitimate.
5076 @findex break_out_memory_refs
5077 @var{x} will always be the result of a call to @code{break_out_memory_refs},
5078 and @var{oldx} will be the operand that was given to that function to produce
5081 The code generated by this macro should not alter the substructure of
5082 @var{x}. If it transforms @var{x} into a more legitimate form, it
5083 should assign @var{x} (which will always be a C variable) a new value.
5085 It is not necessary for this macro to come up with a legitimate
5086 address. The compiler has standard ways of doing so in all cases. In
5087 fact, it is safe to omit this macro. But often a
5088 machine-dependent strategy can generate better code.
5091 @defmac LEGITIMIZE_RELOAD_ADDRESS (@var{x}, @var{mode}, @var{opnum}, @var{type}, @var{ind_levels}, @var{win})
5092 A C compound statement that attempts to replace @var{x}, which is an address
5093 that needs reloading, with a valid memory address for an operand of mode
5094 @var{mode}. @var{win} will be a C statement label elsewhere in the code.
5095 It is not necessary to define this macro, but it might be useful for
5096 performance reasons.
5098 For example, on the i386, it is sometimes possible to use a single
5099 reload register instead of two by reloading a sum of two pseudo
5100 registers into a register. On the other hand, for number of RISC
5101 processors offsets are limited so that often an intermediate address
5102 needs to be generated in order to address a stack slot. By defining
5103 @code{LEGITIMIZE_RELOAD_ADDRESS} appropriately, the intermediate addresses
5104 generated for adjacent some stack slots can be made identical, and thus
5107 @emph{Note}: This macro should be used with caution. It is necessary
5108 to know something of how reload works in order to effectively use this,
5109 and it is quite easy to produce macros that build in too much knowledge
5110 of reload internals.
5112 @emph{Note}: This macro must be able to reload an address created by a
5113 previous invocation of this macro. If it fails to handle such addresses
5114 then the compiler may generate incorrect code or abort.
5117 The macro definition should use @code{push_reload} to indicate parts that
5118 need reloading; @var{opnum}, @var{type} and @var{ind_levels} are usually
5119 suitable to be passed unaltered to @code{push_reload}.
5121 The code generated by this macro must not alter the substructure of
5122 @var{x}. If it transforms @var{x} into a more legitimate form, it
5123 should assign @var{x} (which will always be a C variable) a new value.
5124 This also applies to parts that you change indirectly by calling
5127 @findex strict_memory_address_p
5128 The macro definition may use @code{strict_memory_address_p} to test if
5129 the address has become legitimate.
5132 If you want to change only a part of @var{x}, one standard way of doing
5133 this is to use @code{copy_rtx}. Note, however, that is unshares only a
5134 single level of rtl. Thus, if the part to be changed is not at the
5135 top level, you'll need to replace first the top level.
5136 It is not necessary for this macro to come up with a legitimate
5137 address; but often a machine-dependent strategy can generate better code.
5140 @defmac GO_IF_MODE_DEPENDENT_ADDRESS (@var{addr}, @var{label})
5141 A C statement or compound statement with a conditional @code{goto
5142 @var{label};} executed if memory address @var{x} (an RTX) can have
5143 different meanings depending on the machine mode of the memory
5144 reference it is used for or if the address is valid for some modes
5147 Autoincrement and autodecrement addresses typically have mode-dependent
5148 effects because the amount of the increment or decrement is the size
5149 of the operand being addressed. Some machines have other mode-dependent
5150 addresses. Many RISC machines have no mode-dependent addresses.
5152 You may assume that @var{addr} is a valid address for the machine.
5155 @defmac LEGITIMATE_CONSTANT_P (@var{x})
5156 A C expression that is nonzero if @var{x} is a legitimate constant for
5157 an immediate operand on the target machine. You can assume that
5158 @var{x} satisfies @code{CONSTANT_P}, so you need not check this. In fact,
5159 @samp{1} is a suitable definition for this macro on machines where
5160 anything @code{CONSTANT_P} is valid.
5163 @deftypefn {Target Hook} rtx TARGET_DELEGITIMIZE_ADDRESS (rtx @var{x})
5164 This hook is used to undo the possibly obfuscating effects of the
5165 @code{LEGITIMIZE_ADDRESS} and @code{LEGITIMIZE_RELOAD_ADDRESS} target
5166 macros. Some backend implementations of these macros wrap symbol
5167 references inside an @code{UNSPEC} rtx to represent PIC or similar
5168 addressing modes. This target hook allows GCC's optimizers to understand
5169 the semantics of these opaque @code{UNSPEC}s by converting them back
5170 into their original form.
5173 @deftypefn {Target Hook} bool TARGET_CANNOT_FORCE_CONST_MEM (rtx @var{x})
5174 This hook should return true if @var{x} is of a form that cannot (or
5175 should not) be spilled to the constant pool. The default version of
5176 this hook returns false.
5178 The primary reason to define this hook is to prevent reload from
5179 deciding that a non-legitimate constant would be better reloaded
5180 from the constant pool instead of spilling and reloading a register
5181 holding the constant. This restriction is often true of addresses
5182 of TLS symbols for various targets.
5185 @deftypefn {Target Hook} tree TARGET_VECTORIZE_BUILTIN_MASK_FOR_LOAD (void)
5186 This hook should return the DECL of a function @var{f} that given an
5187 address @var{addr} as an argument returns a mask @var{m} that can be
5188 used to extract from two vectors the relevant data that resides in
5189 @var{addr} in case @var{addr} is not properly aligned.
5191 The autovectrizer, when vectorizing a load operation from an address
5192 @var{addr} that may be unaligned, will generate two vector loads from
5193 the two aligned addresses around @var{addr}. It then generates a
5194 @code{REALIGN_LOAD} operation to extract the relevant data from the
5195 two loaded vectors. The first two arguments to @code{REALIGN_LOAD},
5196 @var{v1} and @var{v2}, are the two vectors, each of size @var{VS}, and
5197 the third argument, @var{OFF}, defines how the data will be extracted
5198 from these two vectors: if @var{OFF} is 0, then the returned vector is
5199 @var{v2}; otherwise, the returned vector is composed from the last
5200 @var{VS}-@var{OFF} elements of @var{v1} concatenated to the first
5201 @var{OFF} elements of @var{v2}.
5203 If this hook is defined, the autovectorizer will generate a call
5204 to @var{f} (using the DECL tree that this hook returns) and will
5205 use the return value of @var{f} as the argument @var{OFF} to
5206 @code{REALIGN_LOAD}. Therefore, the mask @var{m} returned by @var{f}
5207 should comply with the semantics expected by @code{REALIGN_LOAD}
5209 If this hook is not defined, then @var{addr} will be used as
5210 the argument @var{OFF} to @code{REALIGN_LOAD}, in which case the low
5211 log2(@var{VS})-1 bits of @var{addr} will be considered.
5214 @node Condition Code
5215 @section Condition Code Status
5216 @cindex condition code status
5218 @c prevent bad page break with this line
5219 This describes the condition code status.
5222 The file @file{conditions.h} defines a variable @code{cc_status} to
5223 describe how the condition code was computed (in case the interpretation of
5224 the condition code depends on the instruction that it was set by). This
5225 variable contains the RTL expressions on which the condition code is
5226 currently based, and several standard flags.
5228 Sometimes additional machine-specific flags must be defined in the machine
5229 description header file. It can also add additional machine-specific
5230 information by defining @code{CC_STATUS_MDEP}.
5232 @defmac CC_STATUS_MDEP
5233 C code for a data type which is used for declaring the @code{mdep}
5234 component of @code{cc_status}. It defaults to @code{int}.
5236 This macro is not used on machines that do not use @code{cc0}.
5239 @defmac CC_STATUS_MDEP_INIT
5240 A C expression to initialize the @code{mdep} field to ``empty''.
5241 The default definition does nothing, since most machines don't use
5242 the field anyway. If you want to use the field, you should probably
5243 define this macro to initialize it.
5245 This macro is not used on machines that do not use @code{cc0}.
5248 @defmac NOTICE_UPDATE_CC (@var{exp}, @var{insn})
5249 A C compound statement to set the components of @code{cc_status}
5250 appropriately for an insn @var{insn} whose body is @var{exp}. It is
5251 this macro's responsibility to recognize insns that set the condition
5252 code as a byproduct of other activity as well as those that explicitly
5255 This macro is not used on machines that do not use @code{cc0}.
5257 If there are insns that do not set the condition code but do alter
5258 other machine registers, this macro must check to see whether they
5259 invalidate the expressions that the condition code is recorded as
5260 reflecting. For example, on the 68000, insns that store in address
5261 registers do not set the condition code, which means that usually
5262 @code{NOTICE_UPDATE_CC} can leave @code{cc_status} unaltered for such
5263 insns. But suppose that the previous insn set the condition code
5264 based on location @samp{a4@@(102)} and the current insn stores a new
5265 value in @samp{a4}. Although the condition code is not changed by
5266 this, it will no longer be true that it reflects the contents of
5267 @samp{a4@@(102)}. Therefore, @code{NOTICE_UPDATE_CC} must alter
5268 @code{cc_status} in this case to say that nothing is known about the
5269 condition code value.
5271 The definition of @code{NOTICE_UPDATE_CC} must be prepared to deal
5272 with the results of peephole optimization: insns whose patterns are
5273 @code{parallel} RTXs containing various @code{reg}, @code{mem} or
5274 constants which are just the operands. The RTL structure of these
5275 insns is not sufficient to indicate what the insns actually do. What
5276 @code{NOTICE_UPDATE_CC} should do when it sees one is just to run
5277 @code{CC_STATUS_INIT}.
5279 A possible definition of @code{NOTICE_UPDATE_CC} is to call a function
5280 that looks at an attribute (@pxref{Insn Attributes}) named, for example,
5281 @samp{cc}. This avoids having detailed information about patterns in
5282 two places, the @file{md} file and in @code{NOTICE_UPDATE_CC}.
5285 @defmac SELECT_CC_MODE (@var{op}, @var{x}, @var{y})
5286 Returns a mode from class @code{MODE_CC} to be used when comparison
5287 operation code @var{op} is applied to rtx @var{x} and @var{y}. For
5288 example, on the SPARC, @code{SELECT_CC_MODE} is defined as (see
5289 @pxref{Jump Patterns} for a description of the reason for this
5293 #define SELECT_CC_MODE(OP,X,Y) \
5294 (GET_MODE_CLASS (GET_MODE (X)) == MODE_FLOAT \
5295 ? ((OP == EQ || OP == NE) ? CCFPmode : CCFPEmode) \
5296 : ((GET_CODE (X) == PLUS || GET_CODE (X) == MINUS \
5297 || GET_CODE (X) == NEG) \
5298 ? CC_NOOVmode : CCmode))
5301 You should define this macro if and only if you define extra CC modes
5302 in @file{@var{machine}-modes.def}.
5305 @defmac CANONICALIZE_COMPARISON (@var{code}, @var{op0}, @var{op1})
5306 On some machines not all possible comparisons are defined, but you can
5307 convert an invalid comparison into a valid one. For example, the Alpha
5308 does not have a @code{GT} comparison, but you can use an @code{LT}
5309 comparison instead and swap the order of the operands.
5311 On such machines, define this macro to be a C statement to do any
5312 required conversions. @var{code} is the initial comparison code
5313 and @var{op0} and @var{op1} are the left and right operands of the
5314 comparison, respectively. You should modify @var{code}, @var{op0}, and
5315 @var{op1} as required.
5317 GCC will not assume that the comparison resulting from this macro is
5318 valid but will see if the resulting insn matches a pattern in the
5321 You need not define this macro if it would never change the comparison
5325 @defmac REVERSIBLE_CC_MODE (@var{mode})
5326 A C expression whose value is one if it is always safe to reverse a
5327 comparison whose mode is @var{mode}. If @code{SELECT_CC_MODE}
5328 can ever return @var{mode} for a floating-point inequality comparison,
5329 then @code{REVERSIBLE_CC_MODE (@var{mode})} must be zero.
5331 You need not define this macro if it would always returns zero or if the
5332 floating-point format is anything other than @code{IEEE_FLOAT_FORMAT}.
5333 For example, here is the definition used on the SPARC, where floating-point
5334 inequality comparisons are always given @code{CCFPEmode}:
5337 #define REVERSIBLE_CC_MODE(MODE) ((MODE) != CCFPEmode)
5341 @defmac REVERSE_CONDITION (@var{code}, @var{mode})
5342 A C expression whose value is reversed condition code of the @var{code} for
5343 comparison done in CC_MODE @var{mode}. The macro is used only in case
5344 @code{REVERSIBLE_CC_MODE (@var{mode})} is nonzero. Define this macro in case
5345 machine has some non-standard way how to reverse certain conditionals. For
5346 instance in case all floating point conditions are non-trapping, compiler may
5347 freely convert unordered compares to ordered one. Then definition may look
5351 #define REVERSE_CONDITION(CODE, MODE) \
5352 ((MODE) != CCFPmode ? reverse_condition (CODE) \
5353 : reverse_condition_maybe_unordered (CODE))
5357 @defmac REVERSE_CONDEXEC_PREDICATES_P (@var{op1}, @var{op2})
5358 A C expression that returns true if the conditional execution predicate
5359 @var{op1}, a comparison operation, is the inverse of @var{op2} and vice
5360 versa. Define this to return 0 if the target has conditional execution
5361 predicates that cannot be reversed safely. There is no need to validate
5362 that the arguments of op1 and op2 are the same, this is done separately.
5363 If no expansion is specified, this macro is defined as follows:
5366 #define REVERSE_CONDEXEC_PREDICATES_P (x, y) \
5367 (GET_CODE ((x)) == reversed_comparison_code ((y), NULL))
5371 @deftypefn {Target Hook} bool TARGET_FIXED_CONDITION_CODE_REGS (unsigned int *, unsigned int *)
5372 On targets which do not use @code{(cc0)}, and which use a hard
5373 register rather than a pseudo-register to hold condition codes, the
5374 regular CSE passes are often not able to identify cases in which the
5375 hard register is set to a common value. Use this hook to enable a
5376 small pass which optimizes such cases. This hook should return true
5377 to enable this pass, and it should set the integers to which its
5378 arguments point to the hard register numbers used for condition codes.
5379 When there is only one such register, as is true on most systems, the
5380 integer pointed to by the second argument should be set to
5381 @code{INVALID_REGNUM}.
5383 The default version of this hook returns false.
5386 @deftypefn {Target Hook} enum machine_mode TARGET_CC_MODES_COMPATIBLE (enum machine_mode, enum machine_mode)
5387 On targets which use multiple condition code modes in class
5388 @code{MODE_CC}, it is sometimes the case that a comparison can be
5389 validly done in more than one mode. On such a system, define this
5390 target hook to take two mode arguments and to return a mode in which
5391 both comparisons may be validly done. If there is no such mode,
5392 return @code{VOIDmode}.
5394 The default version of this hook checks whether the modes are the
5395 same. If they are, it returns that mode. If they are different, it
5396 returns @code{VOIDmode}.
5400 @section Describing Relative Costs of Operations
5401 @cindex costs of instructions
5402 @cindex relative costs
5403 @cindex speed of instructions
5405 These macros let you describe the relative speed of various operations
5406 on the target machine.
5408 @defmac REGISTER_MOVE_COST (@var{mode}, @var{from}, @var{to})
5409 A C expression for the cost of moving data of mode @var{mode} from a
5410 register in class @var{from} to one in class @var{to}. The classes are
5411 expressed using the enumeration values such as @code{GENERAL_REGS}. A
5412 value of 2 is the default; other values are interpreted relative to
5415 It is not required that the cost always equal 2 when @var{from} is the
5416 same as @var{to}; on some machines it is expensive to move between
5417 registers if they are not general registers.
5419 If reload sees an insn consisting of a single @code{set} between two
5420 hard registers, and if @code{REGISTER_MOVE_COST} applied to their
5421 classes returns a value of 2, reload does not check to ensure that the
5422 constraints of the insn are met. Setting a cost of other than 2 will
5423 allow reload to verify that the constraints are met. You should do this
5424 if the @samp{mov@var{m}} pattern's constraints do not allow such copying.
5427 @defmac MEMORY_MOVE_COST (@var{mode}, @var{class}, @var{in})
5428 A C expression for the cost of moving data of mode @var{mode} between a
5429 register of class @var{class} and memory; @var{in} is zero if the value
5430 is to be written to memory, nonzero if it is to be read in. This cost
5431 is relative to those in @code{REGISTER_MOVE_COST}. If moving between
5432 registers and memory is more expensive than between two registers, you
5433 should define this macro to express the relative cost.
5435 If you do not define this macro, GCC uses a default cost of 4 plus
5436 the cost of copying via a secondary reload register, if one is
5437 needed. If your machine requires a secondary reload register to copy
5438 between memory and a register of @var{class} but the reload mechanism is
5439 more complex than copying via an intermediate, define this macro to
5440 reflect the actual cost of the move.
5442 GCC defines the function @code{memory_move_secondary_cost} if
5443 secondary reloads are needed. It computes the costs due to copying via
5444 a secondary register. If your machine copies from memory using a
5445 secondary register in the conventional way but the default base value of
5446 4 is not correct for your machine, define this macro to add some other
5447 value to the result of that function. The arguments to that function
5448 are the same as to this macro.
5452 A C expression for the cost of a branch instruction. A value of 1 is
5453 the default; other values are interpreted relative to that.
5456 Here are additional macros which do not specify precise relative costs,
5457 but only that certain actions are more expensive than GCC would
5460 @defmac SLOW_BYTE_ACCESS
5461 Define this macro as a C expression which is nonzero if accessing less
5462 than a word of memory (i.e.@: a @code{char} or a @code{short}) is no
5463 faster than accessing a word of memory, i.e., if such access
5464 require more than one instruction or if there is no difference in cost
5465 between byte and (aligned) word loads.
5467 When this macro is not defined, the compiler will access a field by
5468 finding the smallest containing object; when it is defined, a fullword
5469 load will be used if alignment permits. Unless bytes accesses are
5470 faster than word accesses, using word accesses is preferable since it
5471 may eliminate subsequent memory access if subsequent accesses occur to
5472 other fields in the same word of the structure, but to different bytes.
5475 @defmac SLOW_UNALIGNED_ACCESS (@var{mode}, @var{alignment})
5476 Define this macro to be the value 1 if memory accesses described by the
5477 @var{mode} and @var{alignment} parameters have a cost many times greater
5478 than aligned accesses, for example if they are emulated in a trap
5481 When this macro is nonzero, the compiler will act as if
5482 @code{STRICT_ALIGNMENT} were nonzero when generating code for block
5483 moves. This can cause significantly more instructions to be produced.
5484 Therefore, do not set this macro nonzero if unaligned accesses only add a
5485 cycle or two to the time for a memory access.
5487 If the value of this macro is always zero, it need not be defined. If
5488 this macro is defined, it should produce a nonzero value when
5489 @code{STRICT_ALIGNMENT} is nonzero.
5493 The threshold of number of scalar memory-to-memory move insns, @emph{below}
5494 which a sequence of insns should be generated instead of a
5495 string move insn or a library call. Increasing the value will always
5496 make code faster, but eventually incurs high cost in increased code size.
5498 Note that on machines where the corresponding move insn is a
5499 @code{define_expand} that emits a sequence of insns, this macro counts
5500 the number of such sequences.
5502 If you don't define this, a reasonable default is used.
5505 @defmac MOVE_BY_PIECES_P (@var{size}, @var{alignment})
5506 A C expression used to determine whether @code{move_by_pieces} will be used to
5507 copy a chunk of memory, or whether some other block move mechanism
5508 will be used. Defaults to 1 if @code{move_by_pieces_ninsns} returns less
5509 than @code{MOVE_RATIO}.
5512 @defmac MOVE_MAX_PIECES
5513 A C expression used by @code{move_by_pieces} to determine the largest unit
5514 a load or store used to copy memory is. Defaults to @code{MOVE_MAX}.
5518 The threshold of number of scalar move insns, @emph{below} which a sequence
5519 of insns should be generated to clear memory instead of a string clear insn
5520 or a library call. Increasing the value will always make code faster, but
5521 eventually incurs high cost in increased code size.
5523 If you don't define this, a reasonable default is used.
5526 @defmac CLEAR_BY_PIECES_P (@var{size}, @var{alignment})
5527 A C expression used to determine whether @code{clear_by_pieces} will be used
5528 to clear a chunk of memory, or whether some other block clear mechanism
5529 will be used. Defaults to 1 if @code{move_by_pieces_ninsns} returns less
5530 than @code{CLEAR_RATIO}.
5533 @defmac STORE_BY_PIECES_P (@var{size}, @var{alignment})
5534 A C expression used to determine whether @code{store_by_pieces} will be
5535 used to set a chunk of memory to a constant value, or whether some other
5536 mechanism will be used. Used by @code{__builtin_memset} when storing
5537 values other than constant zero and by @code{__builtin_strcpy} when
5538 when called with a constant source string.
5539 Defaults to 1 if @code{move_by_pieces_ninsns} returns less
5540 than @code{MOVE_RATIO}.
5543 @defmac USE_LOAD_POST_INCREMENT (@var{mode})
5544 A C expression used to determine whether a load postincrement is a good
5545 thing to use for a given mode. Defaults to the value of
5546 @code{HAVE_POST_INCREMENT}.
5549 @defmac USE_LOAD_POST_DECREMENT (@var{mode})
5550 A C expression used to determine whether a load postdecrement is a good
5551 thing to use for a given mode. Defaults to the value of
5552 @code{HAVE_POST_DECREMENT}.
5555 @defmac USE_LOAD_PRE_INCREMENT (@var{mode})
5556 A C expression used to determine whether a load preincrement is a good
5557 thing to use for a given mode. Defaults to the value of
5558 @code{HAVE_PRE_INCREMENT}.
5561 @defmac USE_LOAD_PRE_DECREMENT (@var{mode})
5562 A C expression used to determine whether a load predecrement is a good
5563 thing to use for a given mode. Defaults to the value of
5564 @code{HAVE_PRE_DECREMENT}.
5567 @defmac USE_STORE_POST_INCREMENT (@var{mode})
5568 A C expression used to determine whether a store postincrement is a good
5569 thing to use for a given mode. Defaults to the value of
5570 @code{HAVE_POST_INCREMENT}.
5573 @defmac USE_STORE_POST_DECREMENT (@var{mode})
5574 A C expression used to determine whether a store postdecrement is a good
5575 thing to use for a given mode. Defaults to the value of
5576 @code{HAVE_POST_DECREMENT}.
5579 @defmac USE_STORE_PRE_INCREMENT (@var{mode})
5580 This macro is used to determine whether a store preincrement is a good
5581 thing to use for a given mode. Defaults to the value of
5582 @code{HAVE_PRE_INCREMENT}.
5585 @defmac USE_STORE_PRE_DECREMENT (@var{mode})
5586 This macro is used to determine whether a store predecrement is a good
5587 thing to use for a given mode. Defaults to the value of
5588 @code{HAVE_PRE_DECREMENT}.
5591 @defmac NO_FUNCTION_CSE
5592 Define this macro if it is as good or better to call a constant
5593 function address than to call an address kept in a register.
5596 @defmac RANGE_TEST_NON_SHORT_CIRCUIT
5597 Define this macro if a non-short-circuit operation produced by
5598 @samp{fold_range_test ()} is optimal. This macro defaults to true if
5599 @code{BRANCH_COST} is greater than or equal to the value 2.
5602 @deftypefn {Target Hook} bool TARGET_RTX_COSTS (rtx @var{x}, int @var{code}, int @var{outer_code}, int *@var{total})
5603 This target hook describes the relative costs of RTL expressions.
5605 The cost may depend on the precise form of the expression, which is
5606 available for examination in @var{x}, and the rtx code of the expression
5607 in which it is contained, found in @var{outer_code}. @var{code} is the
5608 expression code---redundant, since it can be obtained with
5609 @code{GET_CODE (@var{x})}.
5611 In implementing this hook, you can use the construct
5612 @code{COSTS_N_INSNS (@var{n})} to specify a cost equal to @var{n} fast
5615 On entry to the hook, @code{*@var{total}} contains a default estimate
5616 for the cost of the expression. The hook should modify this value as
5617 necessary. Traditionally, the default costs are @code{COSTS_N_INSNS (5)}
5618 for multiplications, @code{COSTS_N_INSNS (7)} for division and modulus
5619 operations, and @code{COSTS_N_INSNS (1)} for all other operations.
5621 When optimizing for code size, i.e.@: when @code{optimize_size} is
5622 nonzero, this target hook should be used to estimate the relative
5623 size cost of an expression, again relative to @code{COSTS_N_INSNS}.
5625 The hook returns true when all subexpressions of @var{x} have been
5626 processed, and false when @code{rtx_cost} should recurse.
5629 @deftypefn {Target Hook} int TARGET_ADDRESS_COST (rtx @var{address})
5630 This hook computes the cost of an addressing mode that contains
5631 @var{address}. If not defined, the cost is computed from
5632 the @var{address} expression and the @code{TARGET_RTX_COST} hook.
5634 For most CISC machines, the default cost is a good approximation of the
5635 true cost of the addressing mode. However, on RISC machines, all
5636 instructions normally have the same length and execution time. Hence
5637 all addresses will have equal costs.
5639 In cases where more than one form of an address is known, the form with
5640 the lowest cost will be used. If multiple forms have the same, lowest,
5641 cost, the one that is the most complex will be used.
5643 For example, suppose an address that is equal to the sum of a register
5644 and a constant is used twice in the same basic block. When this macro
5645 is not defined, the address will be computed in a register and memory
5646 references will be indirect through that register. On machines where
5647 the cost of the addressing mode containing the sum is no higher than
5648 that of a simple indirect reference, this will produce an additional
5649 instruction and possibly require an additional register. Proper
5650 specification of this macro eliminates this overhead for such machines.
5652 This hook is never called with an invalid address.
5654 On machines where an address involving more than one register is as
5655 cheap as an address computation involving only one register, defining
5656 @code{TARGET_ADDRESS_COST} to reflect this can cause two registers to
5657 be live over a region of code where only one would have been if
5658 @code{TARGET_ADDRESS_COST} were not defined in that manner. This effect
5659 should be considered in the definition of this macro. Equivalent costs
5660 should probably only be given to addresses with different numbers of
5661 registers on machines with lots of registers.
5665 @section Adjusting the Instruction Scheduler
5667 The instruction scheduler may need a fair amount of machine-specific
5668 adjustment in order to produce good code. GCC provides several target
5669 hooks for this purpose. It is usually enough to define just a few of
5670 them: try the first ones in this list first.
5672 @deftypefn {Target Hook} int TARGET_SCHED_ISSUE_RATE (void)
5673 This hook returns the maximum number of instructions that can ever
5674 issue at the same time on the target machine. The default is one.
5675 Although the insn scheduler can define itself the possibility of issue
5676 an insn on the same cycle, the value can serve as an additional
5677 constraint to issue insns on the same simulated processor cycle (see
5678 hooks @samp{TARGET_SCHED_REORDER} and @samp{TARGET_SCHED_REORDER2}).
5679 This value must be constant over the entire compilation. If you need
5680 it to vary depending on what the instructions are, you must use
5681 @samp{TARGET_SCHED_VARIABLE_ISSUE}.
5683 You could define this hook to return the value of the macro
5684 @code{MAX_DFA_ISSUE_RATE}.
5687 @deftypefn {Target Hook} int TARGET_SCHED_VARIABLE_ISSUE (FILE *@var{file}, int @var{verbose}, rtx @var{insn}, int @var{more})
5688 This hook is executed by the scheduler after it has scheduled an insn
5689 from the ready list. It should return the number of insns which can
5690 still be issued in the current cycle. The default is
5691 @samp{@w{@var{more} - 1}} for insns other than @code{CLOBBER} and
5692 @code{USE}, which normally are not counted against the issue rate.
5693 You should define this hook if some insns take more machine resources
5694 than others, so that fewer insns can follow them in the same cycle.
5695 @var{file} is either a null pointer, or a stdio stream to write any
5696 debug output to. @var{verbose} is the verbose level provided by
5697 @option{-fsched-verbose-@var{n}}. @var{insn} is the instruction that
5701 @deftypefn {Target Hook} int TARGET_SCHED_ADJUST_COST (rtx @var{insn}, rtx @var{link}, rtx @var{dep_insn}, int @var{cost})
5702 This function corrects the value of @var{cost} based on the
5703 relationship between @var{insn} and @var{dep_insn} through the
5704 dependence @var{link}. It should return the new value. The default
5705 is to make no adjustment to @var{cost}. This can be used for example
5706 to specify to the scheduler using the traditional pipeline description
5707 that an output- or anti-dependence does not incur the same cost as a
5708 data-dependence. If the scheduler using the automaton based pipeline
5709 description, the cost of anti-dependence is zero and the cost of
5710 output-dependence is maximum of one and the difference of latency
5711 times of the first and the second insns. If these values are not
5712 acceptable, you could use the hook to modify them too. See also
5713 @pxref{Processor pipeline description}.
5716 @deftypefn {Target Hook} int TARGET_SCHED_ADJUST_PRIORITY (rtx @var{insn}, int @var{priority})
5717 This hook adjusts the integer scheduling priority @var{priority} of
5718 @var{insn}. It should return the new priority. Reduce the priority to
5719 execute @var{insn} earlier, increase the priority to execute @var{insn}
5720 later. Do not define this hook if you do not need to adjust the
5721 scheduling priorities of insns.
5724 @deftypefn {Target Hook} int TARGET_SCHED_REORDER (FILE *@var{file}, int @var{verbose}, rtx *@var{ready}, int *@var{n_readyp}, int @var{clock})
5725 This hook is executed by the scheduler after it has scheduled the ready
5726 list, to allow the machine description to reorder it (for example to
5727 combine two small instructions together on @samp{VLIW} machines).
5728 @var{file} is either a null pointer, or a stdio stream to write any
5729 debug output to. @var{verbose} is the verbose level provided by
5730 @option{-fsched-verbose-@var{n}}. @var{ready} is a pointer to the ready
5731 list of instructions that are ready to be scheduled. @var{n_readyp} is
5732 a pointer to the number of elements in the ready list. The scheduler
5733 reads the ready list in reverse order, starting with
5734 @var{ready}[@var{*n_readyp}-1] and going to @var{ready}[0]. @var{clock}
5735 is the timer tick of the scheduler. You may modify the ready list and
5736 the number of ready insns. The return value is the number of insns that
5737 can issue this cycle; normally this is just @code{issue_rate}. See also
5738 @samp{TARGET_SCHED_REORDER2}.
5741 @deftypefn {Target Hook} int TARGET_SCHED_REORDER2 (FILE *@var{file}, int @var{verbose}, rtx *@var{ready}, int *@var{n_ready}, @var{clock})
5742 Like @samp{TARGET_SCHED_REORDER}, but called at a different time. That
5743 function is called whenever the scheduler starts a new cycle. This one
5744 is called once per iteration over a cycle, immediately after
5745 @samp{TARGET_SCHED_VARIABLE_ISSUE}; it can reorder the ready list and
5746 return the number of insns to be scheduled in the same cycle. Defining
5747 this hook can be useful if there are frequent situations where
5748 scheduling one insn causes other insns to become ready in the same
5749 cycle. These other insns can then be taken into account properly.
5752 @deftypefn {Target Hook} void TARGET_SCHED_DEPENDENCIES_EVALUATION_HOOK (rtx @var{head}, rtx @var{tail})
5753 This hook is called after evaluation forward dependencies of insns in
5754 chain given by two parameter values (@var{head} and @var{tail}
5755 correspondingly) but before insns scheduling of the insn chain. For
5756 example, it can be used for better insn classification if it requires
5757 analysis of dependencies. This hook can use backward and forward
5758 dependencies of the insn scheduler because they are already
5762 @deftypefn {Target Hook} void TARGET_SCHED_INIT (FILE *@var{file}, int @var{verbose}, int @var{max_ready})
5763 This hook is executed by the scheduler at the beginning of each block of
5764 instructions that are to be scheduled. @var{file} is either a null
5765 pointer, or a stdio stream to write any debug output to. @var{verbose}
5766 is the verbose level provided by @option{-fsched-verbose-@var{n}}.
5767 @var{max_ready} is the maximum number of insns in the current scheduling
5768 region that can be live at the same time. This can be used to allocate
5769 scratch space if it is needed, e.g.@: by @samp{TARGET_SCHED_REORDER}.
5772 @deftypefn {Target Hook} void TARGET_SCHED_FINISH (FILE *@var{file}, int @var{verbose})
5773 This hook is executed by the scheduler at the end of each block of
5774 instructions that are to be scheduled. It can be used to perform
5775 cleanup of any actions done by the other scheduling hooks. @var{file}
5776 is either a null pointer, or a stdio stream to write any debug output
5777 to. @var{verbose} is the verbose level provided by
5778 @option{-fsched-verbose-@var{n}}.
5781 @deftypefn {Target Hook} void TARGET_SCHED_INIT_GLOBAL (FILE *@var{file}, int @var{verbose}, int @var{old_max_uid})
5782 This hook is executed by the scheduler after function level initializations.
5783 @var{file} is either a null pointer, or a stdio stream to write any debug output to.
5784 @var{verbose} is the verbose level provided by @option{-fsched-verbose-@var{n}}.
5785 @var{old_max_uid} is the maximum insn uid when scheduling begins.
5788 @deftypefn {Target Hook} void TARGET_SCHED_FINISH_GLOBAL (FILE *@var{file}, int @var{verbose})
5789 This is the cleanup hook corresponding to @code{TARGET_SCHED_INIT_GLOBAL}.
5790 @var{file} is either a null pointer, or a stdio stream to write any debug output to.
5791 @var{verbose} is the verbose level provided by @option{-fsched-verbose-@var{n}}.
5794 @deftypefn {Target Hook} int TARGET_SCHED_DFA_PRE_CYCLE_INSN (void)
5795 The hook returns an RTL insn. The automaton state used in the
5796 pipeline hazard recognizer is changed as if the insn were scheduled
5797 when the new simulated processor cycle starts. Usage of the hook may
5798 simplify the automaton pipeline description for some @acronym{VLIW}
5799 processors. If the hook is defined, it is used only for the automaton
5800 based pipeline description. The default is not to change the state
5801 when the new simulated processor cycle starts.
5804 @deftypefn {Target Hook} void TARGET_SCHED_INIT_DFA_PRE_CYCLE_INSN (void)
5805 The hook can be used to initialize data used by the previous hook.
5808 @deftypefn {Target Hook} int TARGET_SCHED_DFA_POST_CYCLE_INSN (void)
5809 The hook is analogous to @samp{TARGET_SCHED_DFA_PRE_CYCLE_INSN} but used
5810 to changed the state as if the insn were scheduled when the new
5811 simulated processor cycle finishes.
5814 @deftypefn {Target Hook} void TARGET_SCHED_INIT_DFA_POST_CYCLE_INSN (void)
5815 The hook is analogous to @samp{TARGET_SCHED_INIT_DFA_PRE_CYCLE_INSN} but
5816 used to initialize data used by the previous hook.
5819 @deftypefn {Target Hook} int TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD (void)
5820 This hook controls better choosing an insn from the ready insn queue
5821 for the @acronym{DFA}-based insn scheduler. Usually the scheduler
5822 chooses the first insn from the queue. If the hook returns a positive
5823 value, an additional scheduler code tries all permutations of
5824 @samp{TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD ()}
5825 subsequent ready insns to choose an insn whose issue will result in
5826 maximal number of issued insns on the same cycle. For the
5827 @acronym{VLIW} processor, the code could actually solve the problem of
5828 packing simple insns into the @acronym{VLIW} insn. Of course, if the
5829 rules of @acronym{VLIW} packing are described in the automaton.
5831 This code also could be used for superscalar @acronym{RISC}
5832 processors. Let us consider a superscalar @acronym{RISC} processor
5833 with 3 pipelines. Some insns can be executed in pipelines @var{A} or
5834 @var{B}, some insns can be executed only in pipelines @var{B} or
5835 @var{C}, and one insn can be executed in pipeline @var{B}. The
5836 processor may issue the 1st insn into @var{A} and the 2nd one into
5837 @var{B}. In this case, the 3rd insn will wait for freeing @var{B}
5838 until the next cycle. If the scheduler issues the 3rd insn the first,
5839 the processor could issue all 3 insns per cycle.
5841 Actually this code demonstrates advantages of the automaton based
5842 pipeline hazard recognizer. We try quickly and easy many insn
5843 schedules to choose the best one.
5845 The default is no multipass scheduling.
5848 @deftypefn {Target Hook} int TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD (rtx)
5850 This hook controls what insns from the ready insn queue will be
5851 considered for the multipass insn scheduling. If the hook returns
5852 zero for insn passed as the parameter, the insn will be not chosen to
5855 The default is that any ready insns can be chosen to be issued.
5858 @deftypefn {Target Hook} int TARGET_SCHED_DFA_NEW_CYCLE (FILE *, int, rtx, int, int, int *)
5860 This hook is called by the insn scheduler before issuing insn passed
5861 as the third parameter on given cycle. If the hook returns nonzero,
5862 the insn is not issued on given processors cycle. Instead of that,
5863 the processor cycle is advanced. If the value passed through the last
5864 parameter is zero, the insn ready queue is not sorted on the new cycle
5865 start as usually. The first parameter passes file for debugging
5866 output. The second one passes the scheduler verbose level of the
5867 debugging output. The forth and the fifth parameter values are
5868 correspondingly processor cycle on which the previous insn has been
5869 issued and the current processor cycle.
5872 @deftypefn {Target Hook} bool TARGET_SCHED_IS_COSTLY_DEPENDENCE (rtx @var{insn1}, rtx @var{insn2}, rtx @var{dep_link}, int @var{dep_cost}, int @var{distance})
5873 This hook is used to define which dependences are considered costly by
5874 the target, so costly that it is not advisable to schedule the insns that
5875 are involved in the dependence too close to one another. The parameters
5876 to this hook are as follows: The second parameter @var{insn2} is dependent
5877 upon the first parameter @var{insn1}. The dependence between @var{insn1}
5878 and @var{insn2} is represented by the third parameter @var{dep_link}. The
5879 fourth parameter @var{cost} is the cost of the dependence, and the fifth
5880 parameter @var{distance} is the distance in cycles between the two insns.
5881 The hook returns @code{true} if considering the distance between the two
5882 insns the dependence between them is considered costly by the target,
5883 and @code{false} otherwise.
5885 Defining this hook can be useful in multiple-issue out-of-order machines,
5886 where (a) it's practically hopeless to predict the actual data/resource
5887 delays, however: (b) there's a better chance to predict the actual grouping
5888 that will be formed, and (c) correctly emulating the grouping can be very
5889 important. In such targets one may want to allow issuing dependent insns
5890 closer to one another---i.e., closer than the dependence distance; however,
5891 not in cases of "costly dependences", which this hooks allows to define.
5894 Macros in the following table are generated by the program
5895 @file{genattr} and can be useful for writing the hooks.
5897 @defmac MAX_DFA_ISSUE_RATE
5898 The macro definition is generated in the automaton based pipeline
5899 description interface. Its value is calculated from the automaton
5900 based pipeline description and is equal to maximal number of all insns
5901 described in constructions @samp{define_insn_reservation} which can be
5902 issued on the same processor cycle.
5906 @section Dividing the Output into Sections (Texts, Data, @dots{})
5907 @c the above section title is WAY too long. maybe cut the part between
5908 @c the (...)? --mew 10feb93
5910 An object file is divided into sections containing different types of
5911 data. In the most common case, there are three sections: the @dfn{text
5912 section}, which holds instructions and read-only data; the @dfn{data
5913 section}, which holds initialized writable data; and the @dfn{bss
5914 section}, which holds uninitialized data. Some systems have other kinds
5917 The compiler must tell the assembler when to switch sections. These
5918 macros control what commands to output to tell the assembler this. You
5919 can also define additional sections.
5921 @defmac TEXT_SECTION_ASM_OP
5922 A C expression whose value is a string, including spacing, containing the
5923 assembler operation that should precede instructions and read-only data.
5924 Normally @code{"\t.text"} is right.
5927 @defmac HOT_TEXT_SECTION_NAME
5928 If defined, a C string constant for the name of the section containing most
5929 frequently executed functions of the program. If not defined, GCC will provide
5930 a default definition if the target supports named sections.
5933 @defmac UNLIKELY_EXECUTED_TEXT_SECTION_NAME
5934 If defined, a C string constant for the name of the section containing unlikely
5935 executed functions in the program.
5938 @defmac DATA_SECTION_ASM_OP
5939 A C expression whose value is a string, including spacing, containing the
5940 assembler operation to identify the following data as writable initialized
5941 data. Normally @code{"\t.data"} is right.
5944 @defmac READONLY_DATA_SECTION_ASM_OP
5945 A C expression whose value is a string, including spacing, containing the
5946 assembler operation to identify the following data as read-only initialized
5950 @defmac READONLY_DATA_SECTION
5951 A macro naming a function to call to switch to the proper section for
5952 read-only data. The default is to use @code{READONLY_DATA_SECTION_ASM_OP}
5953 if defined, else fall back to @code{text_section}.
5955 The most common definition will be @code{data_section}, if the target
5956 does not have a special read-only data section, and does not put data
5957 in the text section.
5960 @defmac BSS_SECTION_ASM_OP
5961 If defined, a C expression whose value is a string, including spacing,
5962 containing the assembler operation to identify the following data as
5963 uninitialized global data. If not defined, and neither
5964 @code{ASM_OUTPUT_BSS} nor @code{ASM_OUTPUT_ALIGNED_BSS} are defined,
5965 uninitialized global data will be output in the data section if
5966 @option{-fno-common} is passed, otherwise @code{ASM_OUTPUT_COMMON} will be
5970 @defmac INIT_SECTION_ASM_OP
5971 If defined, a C expression whose value is a string, including spacing,
5972 containing the assembler operation to identify the following data as
5973 initialization code. If not defined, GCC will assume such a section does
5977 @defmac FINI_SECTION_ASM_OP
5978 If defined, a C expression whose value is a string, including spacing,
5979 containing the assembler operation to identify the following data as
5980 finalization code. If not defined, GCC will assume such a section does
5984 @defmac INIT_ARRAY_SECTION_ASM_OP
5985 If defined, a C expression whose value is a string, including spacing,
5986 containing the assembler operation to identify the following data as
5987 part of the @code{.init_array} (or equivalent) section. If not
5988 defined, GCC will assume such a section does not exist. Do not define
5989 both this macro and @code{INIT_SECTION_ASM_OP}.
5992 @defmac FINI_ARRAY_SECTION_ASM_OP
5993 If defined, a C expression whose value is a string, including spacing,
5994 containing the assembler operation to identify the following data as
5995 part of the @code{.fini_array} (or equivalent) section. If not
5996 defined, GCC will assume such a section does not exist. Do not define
5997 both this macro and @code{FINI_SECTION_ASM_OP}.
6000 @defmac CRT_CALL_STATIC_FUNCTION (@var{section_op}, @var{function})
6001 If defined, an ASM statement that switches to a different section
6002 via @var{section_op}, calls @var{function}, and switches back to
6003 the text section. This is used in @file{crtstuff.c} if
6004 @code{INIT_SECTION_ASM_OP} or @code{FINI_SECTION_ASM_OP} to calls
6005 to initialization and finalization functions from the init and fini
6006 sections. By default, this macro uses a simple function call. Some
6007 ports need hand-crafted assembly code to avoid dependencies on
6008 registers initialized in the function prologue or to ensure that
6009 constant pools don't end up too far way in the text section.
6012 @defmac FORCE_CODE_SECTION_ALIGN
6013 If defined, an ASM statement that aligns a code section to some
6014 arbitrary boundary. This is used to force all fragments of the
6015 @code{.init} and @code{.fini} sections to have to same alignment
6016 and thus prevent the linker from having to add any padding.
6021 @defmac EXTRA_SECTIONS
6022 A list of names for sections other than the standard two, which are
6023 @code{in_text} and @code{in_data}. You need not define this macro
6024 on a system with no other sections (that GCC needs to use).
6027 @findex text_section
6028 @findex data_section
6029 @defmac EXTRA_SECTION_FUNCTIONS
6030 One or more functions to be defined in @file{varasm.c}. These
6031 functions should do jobs analogous to those of @code{text_section} and
6032 @code{data_section}, for your additional sections. Do not define this
6033 macro if you do not define @code{EXTRA_SECTIONS}.
6036 @defmac JUMP_TABLES_IN_TEXT_SECTION
6037 Define this macro to be an expression with a nonzero value if jump
6038 tables (for @code{tablejump} insns) should be output in the text
6039 section, along with the assembler instructions. Otherwise, the
6040 readonly data section is used.
6042 This macro is irrelevant if there is no separate readonly data section.
6045 @deftypefn {Target Hook} void TARGET_ASM_SELECT_SECTION (tree @var{exp}, int @var{reloc}, unsigned HOST_WIDE_INT @var{align})
6046 Switches to the appropriate section for output of @var{exp}. You can
6047 assume that @var{exp} is either a @code{VAR_DECL} node or a constant of
6048 some sort. @var{reloc} indicates whether the initial value of @var{exp}
6049 requires link-time relocations. Bit 0 is set when variable contains
6050 local relocations only, while bit 1 is set for global relocations.
6051 Select the section by calling @code{data_section} or one of the
6052 alternatives for other sections. @var{align} is the constant alignment
6055 The default version of this function takes care of putting read-only
6056 variables in @code{readonly_data_section}.
6058 See also @var{USE_SELECT_SECTION_FOR_FUNCTIONS}.
6061 @defmac USE_SELECT_SECTION_FOR_FUNCTIONS
6062 Define this macro if you wish TARGET_ASM_SELECT_SECTION to be called
6063 for @code{FUNCTION_DECL}s as well as for variables and constants.
6065 In the case of a @code{FUNCTION_DECL}, @var{reloc} will be zero if the
6066 function has been determined to be likely to be called, and nonzero if
6067 it is unlikely to be called.
6070 @deftypefn {Target Hook} void TARGET_ASM_UNIQUE_SECTION (tree @var{decl}, int @var{reloc})
6071 Build up a unique section name, expressed as a @code{STRING_CST} node,
6072 and assign it to @samp{DECL_SECTION_NAME (@var{decl})}.
6073 As with @code{TARGET_ASM_SELECT_SECTION}, @var{reloc} indicates whether
6074 the initial value of @var{exp} requires link-time relocations.
6076 The default version of this function appends the symbol name to the
6077 ELF section name that would normally be used for the symbol. For
6078 example, the function @code{foo} would be placed in @code{.text.foo}.
6079 Whatever the actual target object format, this is often good enough.
6082 @deftypefn {Target Hook} void TARGET_ASM_FUNCTION_RODATA_SECTION (tree @var{decl})
6083 Switches to a readonly data section associated with
6084 @samp{DECL_SECTION_NAME (@var{decl})}.
6085 The default version of this function switches to @code{.gnu.linkonce.r.name}
6086 section if function's section is @code{.gnu.linkonce.t.name}, to
6087 @code{.rodata.name} if function is in @code{.text.name} section
6088 and otherwise switches to the normal readonly data section.
6091 @deftypefn {Target Hook} void TARGET_ASM_SELECT_RTX_SECTION (enum machine_mode @var{mode}, rtx @var{x}, unsigned HOST_WIDE_INT @var{align})
6092 Switches to the appropriate section for output of constant pool entry
6093 @var{x} in @var{mode}. You can assume that @var{x} is some kind of
6094 constant in RTL@. The argument @var{mode} is redundant except in the
6095 case of a @code{const_int} rtx. Select the section by calling
6096 @code{readonly_data_section} or one of the alternatives for other
6097 sections. @var{align} is the constant alignment in bits.
6099 The default version of this function takes care of putting symbolic
6100 constants in @code{flag_pic} mode in @code{data_section} and everything
6101 else in @code{readonly_data_section}.
6104 @deftypefn {Target Hook} void TARGET_ENCODE_SECTION_INFO (tree @var{decl}, rtx @var{rtl}, int @var{new_decl_p})
6105 Define this hook if references to a symbol or a constant must be
6106 treated differently depending on something about the variable or
6107 function named by the symbol (such as what section it is in).
6109 The hook is executed immediately after rtl has been created for
6110 @var{decl}, which may be a variable or function declaration or
6111 an entry in the constant pool. In either case, @var{rtl} is the
6112 rtl in question. Do @emph{not} use @code{DECL_RTL (@var{decl})}
6113 in this hook; that field may not have been initialized yet.
6115 In the case of a constant, it is safe to assume that the rtl is
6116 a @code{mem} whose address is a @code{symbol_ref}. Most decls
6117 will also have this form, but that is not guaranteed. Global
6118 register variables, for instance, will have a @code{reg} for their
6119 rtl. (Normally the right thing to do with such unusual rtl is
6122 The @var{new_decl_p} argument will be true if this is the first time
6123 that @code{TARGET_ENCODE_SECTION_INFO} has been invoked on this decl. It will
6124 be false for subsequent invocations, which will happen for duplicate
6125 declarations. Whether or not anything must be done for the duplicate
6126 declaration depends on whether the hook examines @code{DECL_ATTRIBUTES}.
6127 @var{new_decl_p} is always true when the hook is called for a constant.
6129 @cindex @code{SYMBOL_REF_FLAG}, in @code{TARGET_ENCODE_SECTION_INFO}
6130 The usual thing for this hook to do is to record flags in the
6131 @code{symbol_ref}, using @code{SYMBOL_REF_FLAG} or @code{SYMBOL_REF_FLAGS}.
6132 Historically, the name string was modified if it was necessary to
6133 encode more than one bit of information, but this practice is now
6134 discouraged; use @code{SYMBOL_REF_FLAGS}.
6136 The default definition of this hook, @code{default_encode_section_info}
6137 in @file{varasm.c}, sets a number of commonly-useful bits in
6138 @code{SYMBOL_REF_FLAGS}. Check whether the default does what you need
6139 before overriding it.
6142 @deftypefn {Target Hook} const char *TARGET_STRIP_NAME_ENCODING (const char *name)
6143 Decode @var{name} and return the real name part, sans
6144 the characters that @code{TARGET_ENCODE_SECTION_INFO}
6148 @deftypefn {Target Hook} bool TARGET_IN_SMALL_DATA_P (tree @var{exp})
6149 Returns true if @var{exp} should be placed into a ``small data'' section.
6150 The default version of this hook always returns false.
6153 @deftypevar {Target Hook} bool TARGET_HAVE_SRODATA_SECTION
6154 Contains the value true if the target places read-only
6155 ``small data'' into a separate section. The default value is false.
6158 @deftypefn {Target Hook} bool TARGET_BINDS_LOCAL_P (tree @var{exp})
6159 Returns true if @var{exp} names an object for which name resolution
6160 rules must resolve to the current ``module'' (dynamic shared library
6161 or executable image).
6163 The default version of this hook implements the name resolution rules
6164 for ELF, which has a looser model of global name binding than other
6165 currently supported object file formats.
6168 @deftypevar {Target Hook} bool TARGET_HAVE_TLS
6169 Contains the value true if the target supports thread-local storage.
6170 The default value is false.
6175 @section Position Independent Code
6176 @cindex position independent code
6179 This section describes macros that help implement generation of position
6180 independent code. Simply defining these macros is not enough to
6181 generate valid PIC; you must also add support to the macros
6182 @code{GO_IF_LEGITIMATE_ADDRESS} and @code{PRINT_OPERAND_ADDRESS}, as
6183 well as @code{LEGITIMIZE_ADDRESS}. You must modify the definition of
6184 @samp{movsi} to do something appropriate when the source operand
6185 contains a symbolic address. You may also need to alter the handling of
6186 switch statements so that they use relative addresses.
6187 @c i rearranged the order of the macros above to try to force one of
6188 @c them to the next line, to eliminate an overfull hbox. --mew 10feb93
6190 @defmac PIC_OFFSET_TABLE_REGNUM
6191 The register number of the register used to address a table of static
6192 data addresses in memory. In some cases this register is defined by a
6193 processor's ``application binary interface'' (ABI)@. When this macro
6194 is defined, RTL is generated for this register once, as with the stack
6195 pointer and frame pointer registers. If this macro is not defined, it
6196 is up to the machine-dependent files to allocate such a register (if
6197 necessary). Note that this register must be fixed when in use (e.g.@:
6198 when @code{flag_pic} is true).
6201 @defmac PIC_OFFSET_TABLE_REG_CALL_CLOBBERED
6202 Define this macro if the register defined by
6203 @code{PIC_OFFSET_TABLE_REGNUM} is clobbered by calls. Do not define
6204 this macro if @code{PIC_OFFSET_TABLE_REGNUM} is not defined.
6207 @defmac LEGITIMATE_PIC_OPERAND_P (@var{x})
6208 A C expression that is nonzero if @var{x} is a legitimate immediate
6209 operand on the target machine when generating position independent code.
6210 You can assume that @var{x} satisfies @code{CONSTANT_P}, so you need not
6211 check this. You can also assume @var{flag_pic} is true, so you need not
6212 check it either. You need not define this macro if all constants
6213 (including @code{SYMBOL_REF}) can be immediate operands when generating
6214 position independent code.
6217 @node Assembler Format
6218 @section Defining the Output Assembler Language
6220 This section describes macros whose principal purpose is to describe how
6221 to write instructions in assembler language---rather than what the
6225 * File Framework:: Structural information for the assembler file.
6226 * Data Output:: Output of constants (numbers, strings, addresses).
6227 * Uninitialized Data:: Output of uninitialized variables.
6228 * Label Output:: Output and generation of labels.
6229 * Initialization:: General principles of initialization
6230 and termination routines.
6231 * Macros for Initialization::
6232 Specific macros that control the handling of
6233 initialization and termination routines.
6234 * Instruction Output:: Output of actual instructions.
6235 * Dispatch Tables:: Output of jump tables.
6236 * Exception Region Output:: Output of exception region code.
6237 * Alignment Output:: Pseudo ops for alignment and skipping data.
6240 @node File Framework
6241 @subsection The Overall Framework of an Assembler File
6242 @cindex assembler format
6243 @cindex output of assembler code
6245 @c prevent bad page break with this line
6246 This describes the overall framework of an assembly file.
6248 @deftypefn {Target Hook} void TARGET_ASM_FILE_START ()
6249 @findex default_file_start
6250 Output to @code{asm_out_file} any text which the assembler expects to
6251 find at the beginning of a file. The default behavior is controlled
6252 by two flags, documented below. Unless your target's assembler is
6253 quite unusual, if you override the default, you should call
6254 @code{default_file_start} at some point in your target hook. This
6255 lets other target files rely on these variables.
6258 @deftypevr {Target Hook} bool TARGET_ASM_FILE_START_APP_OFF
6259 If this flag is true, the text of the macro @code{ASM_APP_OFF} will be
6260 printed as the very first line in the assembly file, unless
6261 @option{-fverbose-asm} is in effect. (If that macro has been defined
6262 to the empty string, this variable has no effect.) With the normal
6263 definition of @code{ASM_APP_OFF}, the effect is to notify the GNU
6264 assembler that it need not bother stripping comments or extra
6265 whitespace from its input. This allows it to work a bit faster.
6267 The default is false. You should not set it to true unless you have
6268 verified that your port does not generate any extra whitespace or
6269 comments that will cause GAS to issue errors in NO_APP mode.
6272 @deftypevr {Target Hook} bool TARGET_ASM_FILE_START_FILE_DIRECTIVE
6273 If this flag is true, @code{output_file_directive} will be called
6274 for the primary source file, immediately after printing
6275 @code{ASM_APP_OFF} (if that is enabled). Most ELF assemblers expect
6276 this to be done. The default is false.
6279 @deftypefn {Target Hook} void TARGET_ASM_FILE_END ()
6280 Output to @code{asm_out_file} any text which the assembler expects
6281 to find at the end of a file. The default is to output nothing.
6284 @deftypefun void file_end_indicate_exec_stack ()
6285 Some systems use a common convention, the @samp{.note.GNU-stack}
6286 special section, to indicate whether or not an object file relies on
6287 the stack being executable. If your system uses this convention, you
6288 should define @code{TARGET_ASM_FILE_END} to this function. If you
6289 need to do other things in that hook, have your hook function call
6293 @defmac ASM_COMMENT_START
6294 A C string constant describing how to begin a comment in the target
6295 assembler language. The compiler assumes that the comment will end at
6296 the end of the line.
6300 A C string constant for text to be output before each @code{asm}
6301 statement or group of consecutive ones. Normally this is
6302 @code{"#APP"}, which is a comment that has no effect on most
6303 assemblers but tells the GNU assembler that it must check the lines
6304 that follow for all valid assembler constructs.
6308 A C string constant for text to be output after each @code{asm}
6309 statement or group of consecutive ones. Normally this is
6310 @code{"#NO_APP"}, which tells the GNU assembler to resume making the
6311 time-saving assumptions that are valid for ordinary compiler output.
6314 @defmac ASM_OUTPUT_SOURCE_FILENAME (@var{stream}, @var{name})
6315 A C statement to output COFF information or DWARF debugging information
6316 which indicates that filename @var{name} is the current source file to
6317 the stdio stream @var{stream}.
6319 This macro need not be defined if the standard form of output
6320 for the file format in use is appropriate.
6323 @defmac OUTPUT_QUOTED_STRING (@var{stream}, @var{string})
6324 A C statement to output the string @var{string} to the stdio stream
6325 @var{stream}. If you do not call the function @code{output_quoted_string}
6326 in your config files, GCC will only call it to output filenames to
6327 the assembler source. So you can use it to canonicalize the format
6328 of the filename using this macro.
6331 @defmac ASM_OUTPUT_IDENT (@var{stream}, @var{string})
6332 A C statement to output something to the assembler file to handle a
6333 @samp{#ident} directive containing the text @var{string}. If this
6334 macro is not defined, nothing is output for a @samp{#ident} directive.
6337 @deftypefn {Target Hook} void TARGET_ASM_NAMED_SECTION (const char *@var{name}, unsigned int @var{flags}, unsigned int @var{align})
6338 Output assembly directives to switch to section @var{name}. The section
6339 should have attributes as specified by @var{flags}, which is a bit mask
6340 of the @code{SECTION_*} flags defined in @file{output.h}. If @var{align}
6341 is nonzero, it contains an alignment in bytes to be used for the section,
6342 otherwise some target default should be used. Only targets that must
6343 specify an alignment within the section directive need pay attention to
6344 @var{align} -- we will still use @code{ASM_OUTPUT_ALIGN}.
6347 @deftypefn {Target Hook} bool TARGET_HAVE_NAMED_SECTIONS
6348 This flag is true if the target supports @code{TARGET_ASM_NAMED_SECTION}.
6351 @deftypefn {Target Hook} {unsigned int} TARGET_SECTION_TYPE_FLAGS (tree @var{decl}, const char *@var{name}, int @var{reloc})
6352 Choose a set of section attributes for use by @code{TARGET_ASM_NAMED_SECTION}
6353 based on a variable or function decl, a section name, and whether or not the
6354 declaration's initializer may contain runtime relocations. @var{decl} may be
6355 null, in which case read-write data should be assumed.
6357 The default version if this function handles choosing code vs data,
6358 read-only vs read-write data, and @code{flag_pic}. You should only
6359 need to override this if your target has special flags that might be
6360 set via @code{__attribute__}.
6365 @subsection Output of Data
6368 @deftypevr {Target Hook} {const char *} TARGET_ASM_BYTE_OP
6369 @deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_HI_OP
6370 @deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_SI_OP
6371 @deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_DI_OP
6372 @deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_TI_OP
6373 @deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_HI_OP
6374 @deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_SI_OP
6375 @deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_DI_OP
6376 @deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_TI_OP
6377 These hooks specify assembly directives for creating certain kinds
6378 of integer object. The @code{TARGET_ASM_BYTE_OP} directive creates a
6379 byte-sized object, the @code{TARGET_ASM_ALIGNED_HI_OP} one creates an
6380 aligned two-byte object, and so on. Any of the hooks may be
6381 @code{NULL}, indicating that no suitable directive is available.
6383 The compiler will print these strings at the start of a new line,
6384 followed immediately by the object's initial value. In most cases,
6385 the string should contain a tab, a pseudo-op, and then another tab.
6388 @deftypefn {Target Hook} bool TARGET_ASM_INTEGER (rtx @var{x}, unsigned int @var{size}, int @var{aligned_p})
6389 The @code{assemble_integer} function uses this hook to output an
6390 integer object. @var{x} is the object's value, @var{size} is its size
6391 in bytes and @var{aligned_p} indicates whether it is aligned. The
6392 function should return @code{true} if it was able to output the
6393 object. If it returns false, @code{assemble_integer} will try to
6394 split the object into smaller parts.
6396 The default implementation of this hook will use the
6397 @code{TARGET_ASM_BYTE_OP} family of strings, returning @code{false}
6398 when the relevant string is @code{NULL}.
6401 @defmac OUTPUT_ADDR_CONST_EXTRA (@var{stream}, @var{x}, @var{fail})
6402 A C statement to recognize @var{rtx} patterns that
6403 @code{output_addr_const} can't deal with, and output assembly code to
6404 @var{stream} corresponding to the pattern @var{x}. This may be used to
6405 allow machine-dependent @code{UNSPEC}s to appear within constants.
6407 If @code{OUTPUT_ADDR_CONST_EXTRA} fails to recognize a pattern, it must
6408 @code{goto fail}, so that a standard error message is printed. If it
6409 prints an error message itself, by calling, for example,
6410 @code{output_operand_lossage}, it may just complete normally.
6413 @defmac ASM_OUTPUT_ASCII (@var{stream}, @var{ptr}, @var{len})
6414 A C statement to output to the stdio stream @var{stream} an assembler
6415 instruction to assemble a string constant containing the @var{len}
6416 bytes at @var{ptr}. @var{ptr} will be a C expression of type
6417 @code{char *} and @var{len} a C expression of type @code{int}.
6419 If the assembler has a @code{.ascii} pseudo-op as found in the
6420 Berkeley Unix assembler, do not define the macro
6421 @code{ASM_OUTPUT_ASCII}.
6424 @defmac ASM_OUTPUT_FDESC (@var{stream}, @var{decl}, @var{n})
6425 A C statement to output word @var{n} of a function descriptor for
6426 @var{decl}. This must be defined if @code{TARGET_VTABLE_USES_DESCRIPTORS}
6427 is defined, and is otherwise unused.
6430 @defmac CONSTANT_POOL_BEFORE_FUNCTION
6431 You may define this macro as a C expression. You should define the
6432 expression to have a nonzero value if GCC should output the constant
6433 pool for a function before the code for the function, or a zero value if
6434 GCC should output the constant pool after the function. If you do
6435 not define this macro, the usual case, GCC will output the constant
6436 pool before the function.
6439 @defmac ASM_OUTPUT_POOL_PROLOGUE (@var{file}, @var{funname}, @var{fundecl}, @var{size})
6440 A C statement to output assembler commands to define the start of the
6441 constant pool for a function. @var{funname} is a string giving
6442 the name of the function. Should the return type of the function
6443 be required, it can be obtained via @var{fundecl}. @var{size}
6444 is the size, in bytes, of the constant pool that will be written
6445 immediately after this call.
6447 If no constant-pool prefix is required, the usual case, this macro need
6451 @defmac ASM_OUTPUT_SPECIAL_POOL_ENTRY (@var{file}, @var{x}, @var{mode}, @var{align}, @var{labelno}, @var{jumpto})
6452 A C statement (with or without semicolon) to output a constant in the
6453 constant pool, if it needs special treatment. (This macro need not do
6454 anything for RTL expressions that can be output normally.)
6456 The argument @var{file} is the standard I/O stream to output the
6457 assembler code on. @var{x} is the RTL expression for the constant to
6458 output, and @var{mode} is the machine mode (in case @var{x} is a
6459 @samp{const_int}). @var{align} is the required alignment for the value
6460 @var{x}; you should output an assembler directive to force this much
6463 The argument @var{labelno} is a number to use in an internal label for
6464 the address of this pool entry. The definition of this macro is
6465 responsible for outputting the label definition at the proper place.
6466 Here is how to do this:
6469 @code{(*targetm.asm_out.internal_label)} (@var{file}, "LC", @var{labelno});
6472 When you output a pool entry specially, you should end with a
6473 @code{goto} to the label @var{jumpto}. This will prevent the same pool
6474 entry from being output a second time in the usual manner.
6476 You need not define this macro if it would do nothing.
6479 @defmac ASM_OUTPUT_POOL_EPILOGUE (@var{file} @var{funname} @var{fundecl} @var{size})
6480 A C statement to output assembler commands to at the end of the constant
6481 pool for a function. @var{funname} is a string giving the name of the
6482 function. Should the return type of the function be required, you can
6483 obtain it via @var{fundecl}. @var{size} is the size, in bytes, of the
6484 constant pool that GCC wrote immediately before this call.
6486 If no constant-pool epilogue is required, the usual case, you need not
6490 @defmac IS_ASM_LOGICAL_LINE_SEPARATOR (@var{C})
6491 Define this macro as a C expression which is nonzero if @var{C} is
6492 used as a logical line separator by the assembler.
6494 If you do not define this macro, the default is that only
6495 the character @samp{;} is treated as a logical line separator.
6498 @deftypevr {Target Hook} {const char *} TARGET_ASM_OPEN_PAREN
6499 @deftypevrx {Target Hook} {const char *} TARGET_ASM_CLOSE_PAREN
6500 These target hooks are C string constants, describing the syntax in the
6501 assembler for grouping arithmetic expressions. If not overridden, they
6502 default to normal parentheses, which is correct for most assemblers.
6505 These macros are provided by @file{real.h} for writing the definitions
6506 of @code{ASM_OUTPUT_DOUBLE} and the like:
6508 @defmac REAL_VALUE_TO_TARGET_SINGLE (@var{x}, @var{l})
6509 @defmacx REAL_VALUE_TO_TARGET_DOUBLE (@var{x}, @var{l})
6510 @defmacx REAL_VALUE_TO_TARGET_LONG_DOUBLE (@var{x}, @var{l})
6511 These translate @var{x}, of type @code{REAL_VALUE_TYPE}, to the target's
6512 floating point representation, and store its bit pattern in the variable
6513 @var{l}. For @code{REAL_VALUE_TO_TARGET_SINGLE}, this variable should
6514 be a simple @code{long int}. For the others, it should be an array of
6515 @code{long int}. The number of elements in this array is determined by
6516 the size of the desired target floating point data type: 32 bits of it
6517 go in each @code{long int} array element. Each array element holds 32
6518 bits of the result, even if @code{long int} is wider than 32 bits on the
6521 The array element values are designed so that you can print them out
6522 using @code{fprintf} in the order they should appear in the target
6526 @node Uninitialized Data
6527 @subsection Output of Uninitialized Variables
6529 Each of the macros in this section is used to do the whole job of
6530 outputting a single uninitialized variable.
6532 @defmac ASM_OUTPUT_COMMON (@var{stream}, @var{name}, @var{size}, @var{rounded})
6533 A C statement (sans semicolon) to output to the stdio stream
6534 @var{stream} the assembler definition of a common-label named
6535 @var{name} whose size is @var{size} bytes. The variable @var{rounded}
6536 is the size rounded up to whatever alignment the caller wants.
6538 Use the expression @code{assemble_name (@var{stream}, @var{name})} to
6539 output the name itself; before and after that, output the additional
6540 assembler syntax for defining the name, and a newline.
6542 This macro controls how the assembler definitions of uninitialized
6543 common global variables are output.
6546 @defmac ASM_OUTPUT_ALIGNED_COMMON (@var{stream}, @var{name}, @var{size}, @var{alignment})
6547 Like @code{ASM_OUTPUT_COMMON} except takes the required alignment as a
6548 separate, explicit argument. If you define this macro, it is used in
6549 place of @code{ASM_OUTPUT_COMMON}, and gives you more flexibility in
6550 handling the required alignment of the variable. The alignment is specified
6551 as the number of bits.
6554 @defmac ASM_OUTPUT_ALIGNED_DECL_COMMON (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{alignment})
6555 Like @code{ASM_OUTPUT_ALIGNED_COMMON} except that @var{decl} of the
6556 variable to be output, if there is one, or @code{NULL_TREE} if there
6557 is no corresponding variable. If you define this macro, GCC will use it
6558 in place of both @code{ASM_OUTPUT_COMMON} and
6559 @code{ASM_OUTPUT_ALIGNED_COMMON}. Define this macro when you need to see
6560 the variable's decl in order to chose what to output.
6563 @defmac ASM_OUTPUT_SHARED_COMMON (@var{stream}, @var{name}, @var{size}, @var{rounded})
6564 If defined, it is similar to @code{ASM_OUTPUT_COMMON}, except that it
6565 is used when @var{name} is shared. If not defined, @code{ASM_OUTPUT_COMMON}
6569 @defmac ASM_OUTPUT_BSS (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{rounded})
6570 A C statement (sans semicolon) to output to the stdio stream
6571 @var{stream} the assembler definition of uninitialized global @var{decl} named
6572 @var{name} whose size is @var{size} bytes. The variable @var{rounded}
6573 is the size rounded up to whatever alignment the caller wants.
6575 Try to use function @code{asm_output_bss} defined in @file{varasm.c} when
6576 defining this macro. If unable, use the expression
6577 @code{assemble_name (@var{stream}, @var{name})} to output the name itself;
6578 before and after that, output the additional assembler syntax for defining
6579 the name, and a newline.
6581 This macro controls how the assembler definitions of uninitialized global
6582 variables are output. This macro exists to properly support languages like
6583 C++ which do not have @code{common} data. However, this macro currently
6584 is not defined for all targets. If this macro and
6585 @code{ASM_OUTPUT_ALIGNED_BSS} are not defined then @code{ASM_OUTPUT_COMMON}
6586 or @code{ASM_OUTPUT_ALIGNED_COMMON} or
6587 @code{ASM_OUTPUT_ALIGNED_DECL_COMMON} is used.
6590 @defmac ASM_OUTPUT_ALIGNED_BSS (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{alignment})
6591 Like @code{ASM_OUTPUT_BSS} except takes the required alignment as a
6592 separate, explicit argument. If you define this macro, it is used in
6593 place of @code{ASM_OUTPUT_BSS}, and gives you more flexibility in
6594 handling the required alignment of the variable. The alignment is specified
6595 as the number of bits.
6597 Try to use function @code{asm_output_aligned_bss} defined in file
6598 @file{varasm.c} when defining this macro.
6601 @defmac ASM_OUTPUT_SHARED_BSS (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{rounded})
6602 If defined, it is similar to @code{ASM_OUTPUT_BSS}, except that it
6603 is used when @var{name} is shared. If not defined, @code{ASM_OUTPUT_BSS}
6607 @defmac ASM_OUTPUT_LOCAL (@var{stream}, @var{name}, @var{size}, @var{rounded})
6608 A C statement (sans semicolon) to output to the stdio stream
6609 @var{stream} the assembler definition of a local-common-label named
6610 @var{name} whose size is @var{size} bytes. The variable @var{rounded}
6611 is the size rounded up to whatever alignment the caller wants.
6613 Use the expression @code{assemble_name (@var{stream}, @var{name})} to
6614 output the name itself; before and after that, output the additional
6615 assembler syntax for defining the name, and a newline.
6617 This macro controls how the assembler definitions of uninitialized
6618 static variables are output.
6621 @defmac ASM_OUTPUT_ALIGNED_LOCAL (@var{stream}, @var{name}, @var{size}, @var{alignment})
6622 Like @code{ASM_OUTPUT_LOCAL} except takes the required alignment as a
6623 separate, explicit argument. If you define this macro, it is used in
6624 place of @code{ASM_OUTPUT_LOCAL}, and gives you more flexibility in
6625 handling the required alignment of the variable. The alignment is specified
6626 as the number of bits.
6629 @defmac ASM_OUTPUT_ALIGNED_DECL_LOCAL (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{alignment})
6630 Like @code{ASM_OUTPUT_ALIGNED_DECL} except that @var{decl} of the
6631 variable to be output, if there is one, or @code{NULL_TREE} if there
6632 is no corresponding variable. If you define this macro, GCC will use it
6633 in place of both @code{ASM_OUTPUT_DECL} and
6634 @code{ASM_OUTPUT_ALIGNED_DECL}. Define this macro when you need to see
6635 the variable's decl in order to chose what to output.
6638 @defmac ASM_OUTPUT_SHARED_LOCAL (@var{stream}, @var{name}, @var{size}, @var{rounded})
6639 If defined, it is similar to @code{ASM_OUTPUT_LOCAL}, except that it
6640 is used when @var{name} is shared. If not defined, @code{ASM_OUTPUT_LOCAL}
6645 @subsection Output and Generation of Labels
6647 @c prevent bad page break with this line
6648 This is about outputting labels.
6650 @findex assemble_name
6651 @defmac ASM_OUTPUT_LABEL (@var{stream}, @var{name})
6652 A C statement (sans semicolon) to output to the stdio stream
6653 @var{stream} the assembler definition of a label named @var{name}.
6654 Use the expression @code{assemble_name (@var{stream}, @var{name})} to
6655 output the name itself; before and after that, output the additional
6656 assembler syntax for defining the name, and a newline. A default
6657 definition of this macro is provided which is correct for most systems.
6660 @findex assemble_name_raw
6661 @defmac ASM_OUTPUT_INTERNAL_LABEL (@var{stream}, @var{name})
6662 Identical to @code{ASM_OUTPUT_LABEL}, except that @var{name} is known
6663 to refer to a compiler-generated label. The default definition uses
6664 @code{assemble_name_raw}, which is like @code{assemble_name} except
6665 that it is more efficient.
6669 A C string containing the appropriate assembler directive to specify the
6670 size of a symbol, without any arguments. On systems that use ELF, the
6671 default (in @file{config/elfos.h}) is @samp{"\t.size\t"}; on other
6672 systems, the default is not to define this macro.
6674 Define this macro only if it is correct to use the default definitions
6675 of @code{ASM_OUTPUT_SIZE_DIRECTIVE} and @code{ASM_OUTPUT_MEASURED_SIZE}
6676 for your system. If you need your own custom definitions of those
6677 macros, or if you do not need explicit symbol sizes at all, do not
6681 @defmac ASM_OUTPUT_SIZE_DIRECTIVE (@var{stream}, @var{name}, @var{size})
6682 A C statement (sans semicolon) to output to the stdio stream
6683 @var{stream} a directive telling the assembler that the size of the
6684 symbol @var{name} is @var{size}. @var{size} is a @code{HOST_WIDE_INT}.
6685 If you define @code{SIZE_ASM_OP}, a default definition of this macro is
6689 @defmac ASM_OUTPUT_MEASURED_SIZE (@var{stream}, @var{name})
6690 A C statement (sans semicolon) to output to the stdio stream
6691 @var{stream} a directive telling the assembler to calculate the size of
6692 the symbol @var{name} by subtracting its address from the current
6695 If you define @code{SIZE_ASM_OP}, a default definition of this macro is
6696 provided. The default assumes that the assembler recognizes a special
6697 @samp{.} symbol as referring to the current address, and can calculate
6698 the difference between this and another symbol. If your assembler does
6699 not recognize @samp{.} or cannot do calculations with it, you will need
6700 to redefine @code{ASM_OUTPUT_MEASURED_SIZE} to use some other technique.
6704 A C string containing the appropriate assembler directive to specify the
6705 type of a symbol, without any arguments. On systems that use ELF, the
6706 default (in @file{config/elfos.h}) is @samp{"\t.type\t"}; on other
6707 systems, the default is not to define this macro.
6709 Define this macro only if it is correct to use the default definition of
6710 @code{ASM_OUTPUT_TYPE_DIRECTIVE} for your system. If you need your own
6711 custom definition of this macro, or if you do not need explicit symbol
6712 types at all, do not define this macro.
6715 @defmac TYPE_OPERAND_FMT
6716 A C string which specifies (using @code{printf} syntax) the format of
6717 the second operand to @code{TYPE_ASM_OP}. On systems that use ELF, the
6718 default (in @file{config/elfos.h}) is @samp{"@@%s"}; on other systems,
6719 the default is not to define this macro.
6721 Define this macro only if it is correct to use the default definition of
6722 @code{ASM_OUTPUT_TYPE_DIRECTIVE} for your system. If you need your own
6723 custom definition of this macro, or if you do not need explicit symbol
6724 types at all, do not define this macro.
6727 @defmac ASM_OUTPUT_TYPE_DIRECTIVE (@var{stream}, @var{type})
6728 A C statement (sans semicolon) to output to the stdio stream
6729 @var{stream} a directive telling the assembler that the type of the
6730 symbol @var{name} is @var{type}. @var{type} is a C string; currently,
6731 that string is always either @samp{"function"} or @samp{"object"}, but
6732 you should not count on this.
6734 If you define @code{TYPE_ASM_OP} and @code{TYPE_OPERAND_FMT}, a default
6735 definition of this macro is provided.
6738 @defmac ASM_DECLARE_FUNCTION_NAME (@var{stream}, @var{name}, @var{decl})
6739 A C statement (sans semicolon) to output to the stdio stream
6740 @var{stream} any text necessary for declaring the name @var{name} of a
6741 function which is being defined. This macro is responsible for
6742 outputting the label definition (perhaps using
6743 @code{ASM_OUTPUT_LABEL}). The argument @var{decl} is the
6744 @code{FUNCTION_DECL} tree node representing the function.
6746 If this macro is not defined, then the function name is defined in the
6747 usual manner as a label (by means of @code{ASM_OUTPUT_LABEL}).
6749 You may wish to use @code{ASM_OUTPUT_TYPE_DIRECTIVE} in the definition
6753 @defmac ASM_DECLARE_FUNCTION_SIZE (@var{stream}, @var{name}, @var{decl})
6754 A C statement (sans semicolon) to output to the stdio stream
6755 @var{stream} any text necessary for declaring the size of a function
6756 which is being defined. The argument @var{name} is the name of the
6757 function. The argument @var{decl} is the @code{FUNCTION_DECL} tree node
6758 representing the function.
6760 If this macro is not defined, then the function size is not defined.
6762 You may wish to use @code{ASM_OUTPUT_MEASURED_SIZE} in the definition
6766 @defmac ASM_DECLARE_OBJECT_NAME (@var{stream}, @var{name}, @var{decl})
6767 A C statement (sans semicolon) to output to the stdio stream
6768 @var{stream} any text necessary for declaring the name @var{name} of an
6769 initialized variable which is being defined. This macro must output the
6770 label definition (perhaps using @code{ASM_OUTPUT_LABEL}). The argument
6771 @var{decl} is the @code{VAR_DECL} tree node representing the variable.
6773 If this macro is not defined, then the variable name is defined in the
6774 usual manner as a label (by means of @code{ASM_OUTPUT_LABEL}).
6776 You may wish to use @code{ASM_OUTPUT_TYPE_DIRECTIVE} and/or
6777 @code{ASM_OUTPUT_SIZE_DIRECTIVE} in the definition of this macro.
6780 @defmac ASM_DECLARE_CONSTANT_NAME (@var{stream}, @var{name}, @var{exp}, @var{size})
6781 A C statement (sans semicolon) to output to the stdio stream
6782 @var{stream} any text necessary for declaring the name @var{name} of a
6783 constant which is being defined. This macro is responsible for
6784 outputting the label definition (perhaps using
6785 @code{ASM_OUTPUT_LABEL}). The argument @var{exp} is the
6786 value of the constant, and @var{size} is the size of the constant
6787 in bytes. @var{name} will be an internal label.
6789 If this macro is not defined, then the @var{name} is defined in the
6790 usual manner as a label (by means of @code{ASM_OUTPUT_LABEL}).
6792 You may wish to use @code{ASM_OUTPUT_TYPE_DIRECTIVE} in the definition
6796 @defmac ASM_DECLARE_REGISTER_GLOBAL (@var{stream}, @var{decl}, @var{regno}, @var{name})
6797 A C statement (sans semicolon) to output to the stdio stream
6798 @var{stream} any text necessary for claiming a register @var{regno}
6799 for a global variable @var{decl} with name @var{name}.
6801 If you don't define this macro, that is equivalent to defining it to do
6805 @defmac ASM_FINISH_DECLARE_OBJECT (@var{stream}, @var{decl}, @var{toplevel}, @var{atend})
6806 A C statement (sans semicolon) to finish up declaring a variable name
6807 once the compiler has processed its initializer fully and thus has had a
6808 chance to determine the size of an array when controlled by an
6809 initializer. This is used on systems where it's necessary to declare
6810 something about the size of the object.
6812 If you don't define this macro, that is equivalent to defining it to do
6815 You may wish to use @code{ASM_OUTPUT_SIZE_DIRECTIVE} and/or
6816 @code{ASM_OUTPUT_MEASURED_SIZE} in the definition of this macro.
6819 @deftypefn {Target Hook} void TARGET_ASM_GLOBALIZE_LABEL (FILE *@var{stream}, const char *@var{name})
6820 This target hook is a function to output to the stdio stream
6821 @var{stream} some commands that will make the label @var{name} global;
6822 that is, available for reference from other files.
6824 The default implementation relies on a proper definition of
6825 @code{GLOBAL_ASM_OP}.
6828 @defmac ASM_WEAKEN_LABEL (@var{stream}, @var{name})
6829 A C statement (sans semicolon) to output to the stdio stream
6830 @var{stream} some commands that will make the label @var{name} weak;
6831 that is, available for reference from other files but only used if
6832 no other definition is available. Use the expression
6833 @code{assemble_name (@var{stream}, @var{name})} to output the name
6834 itself; before and after that, output the additional assembler syntax
6835 for making that name weak, and a newline.
6837 If you don't define this macro or @code{ASM_WEAKEN_DECL}, GCC will not
6838 support weak symbols and you should not define the @code{SUPPORTS_WEAK}
6842 @defmac ASM_WEAKEN_DECL (@var{stream}, @var{decl}, @var{name}, @var{value})
6843 Combines (and replaces) the function of @code{ASM_WEAKEN_LABEL} and
6844 @code{ASM_OUTPUT_WEAK_ALIAS}, allowing access to the associated function
6845 or variable decl. If @var{value} is not @code{NULL}, this C statement
6846 should output to the stdio stream @var{stream} assembler code which
6847 defines (equates) the weak symbol @var{name} to have the value
6848 @var{value}. If @var{value} is @code{NULL}, it should output commands
6849 to make @var{name} weak.
6852 @defmac ASM_OUTPUT_WEAKREF (@var{stream}, @var{decl}, @var{name}, @var{value})
6853 Outputs a directive that enables @var{name} to be used to refer to
6854 symbol @var{value} with weak-symbol semantics. @code{decl} is the
6855 declaration of @code{name}.
6858 @defmac SUPPORTS_WEAK
6859 A C expression which evaluates to true if the target supports weak symbols.
6861 If you don't define this macro, @file{defaults.h} provides a default
6862 definition. If either @code{ASM_WEAKEN_LABEL} or @code{ASM_WEAKEN_DECL}
6863 is defined, the default definition is @samp{1}; otherwise, it is
6864 @samp{0}. Define this macro if you want to control weak symbol support
6865 with a compiler flag such as @option{-melf}.
6868 @defmac MAKE_DECL_ONE_ONLY (@var{decl})
6869 A C statement (sans semicolon) to mark @var{decl} to be emitted as a
6870 public symbol such that extra copies in multiple translation units will
6871 be discarded by the linker. Define this macro if your object file
6872 format provides support for this concept, such as the @samp{COMDAT}
6873 section flags in the Microsoft Windows PE/COFF format, and this support
6874 requires changes to @var{decl}, such as putting it in a separate section.
6877 @defmac SUPPORTS_ONE_ONLY
6878 A C expression which evaluates to true if the target supports one-only
6881 If you don't define this macro, @file{varasm.c} provides a default
6882 definition. If @code{MAKE_DECL_ONE_ONLY} is defined, the default
6883 definition is @samp{1}; otherwise, it is @samp{0}. Define this macro if
6884 you want to control one-only symbol support with a compiler flag, or if
6885 setting the @code{DECL_ONE_ONLY} flag is enough to mark a declaration to
6886 be emitted as one-only.
6889 @deftypefn {Target Hook} void TARGET_ASM_ASSEMBLE_VISIBILITY (tree @var{decl}, const char *@var{visibility})
6890 This target hook is a function to output to @var{asm_out_file} some
6891 commands that will make the symbol(s) associated with @var{decl} have
6892 hidden, protected or internal visibility as specified by @var{visibility}.
6895 @defmac TARGET_WEAK_NOT_IN_ARCHIVE_TOC
6896 A C expression that evaluates to true if the target's linker expects
6897 that weak symbols do not appear in a static archive's table of contents.
6898 The default is @code{0}.
6900 Leaving weak symbols out of an archive's table of contents means that,
6901 if a symbol will only have a definition in one translation unit and
6902 will have undefined references from other translation units, that
6903 symbol should not be weak. Defining this macro to be nonzero will
6904 thus have the effect that certain symbols that would normally be weak
6905 (explicit template instantiations, and vtables for polymorphic classes
6906 with noninline key methods) will instead be nonweak.
6908 The C++ ABI requires this macro to be zero. Define this macro for
6909 targets where full C++ ABI compliance is impossible and where linker
6910 restrictions require weak symbols to be left out of a static archive's
6914 @defmac ASM_OUTPUT_EXTERNAL (@var{stream}, @var{decl}, @var{name})
6915 A C statement (sans semicolon) to output to the stdio stream
6916 @var{stream} any text necessary for declaring the name of an external
6917 symbol named @var{name} which is referenced in this compilation but
6918 not defined. The value of @var{decl} is the tree node for the
6921 This macro need not be defined if it does not need to output anything.
6922 The GNU assembler and most Unix assemblers don't require anything.
6925 @deftypefn {Target Hook} void TARGET_ASM_EXTERNAL_LIBCALL (rtx @var{symref})
6926 This target hook is a function to output to @var{asm_out_file} an assembler
6927 pseudo-op to declare a library function name external. The name of the
6928 library function is given by @var{symref}, which is a @code{symbol_ref}.
6931 @deftypefn {Target Hook} void TARGET_ASM_MARK_DECL_PRESERVED (tree @var{decl})
6932 This target hook is a function to output to @var{asm_out_file} an assembler
6933 directive to annotate used symbol. Darwin target use .no_dead_code_strip
6937 @defmac ASM_OUTPUT_LABELREF (@var{stream}, @var{name})
6938 A C statement (sans semicolon) to output to the stdio stream
6939 @var{stream} a reference in assembler syntax to a label named
6940 @var{name}. This should add @samp{_} to the front of the name, if that
6941 is customary on your operating system, as it is in most Berkeley Unix
6942 systems. This macro is used in @code{assemble_name}.
6945 @defmac ASM_OUTPUT_SYMBOL_REF (@var{stream}, @var{sym})
6946 A C statement (sans semicolon) to output a reference to
6947 @code{SYMBOL_REF} @var{sym}. If not defined, @code{assemble_name}
6948 will be used to output the name of the symbol. This macro may be used
6949 to modify the way a symbol is referenced depending on information
6950 encoded by @code{TARGET_ENCODE_SECTION_INFO}.
6953 @defmac ASM_OUTPUT_LABEL_REF (@var{stream}, @var{buf})
6954 A C statement (sans semicolon) to output a reference to @var{buf}, the
6955 result of @code{ASM_GENERATE_INTERNAL_LABEL}. If not defined,
6956 @code{assemble_name} will be used to output the name of the symbol.
6957 This macro is not used by @code{output_asm_label}, or the @code{%l}
6958 specifier that calls it; the intention is that this macro should be set
6959 when it is necessary to output a label differently when its address is
6963 @deftypefn {Target Hook} void TARGET_ASM_INTERNAL_LABEL (FILE *@var{stream}, const char *@var{prefix}, unsigned long @var{labelno})
6964 A function to output to the stdio stream @var{stream} a label whose
6965 name is made from the string @var{prefix} and the number @var{labelno}.
6967 It is absolutely essential that these labels be distinct from the labels
6968 used for user-level functions and variables. Otherwise, certain programs
6969 will have name conflicts with internal labels.
6971 It is desirable to exclude internal labels from the symbol table of the
6972 object file. Most assemblers have a naming convention for labels that
6973 should be excluded; on many systems, the letter @samp{L} at the
6974 beginning of a label has this effect. You should find out what
6975 convention your system uses, and follow it.
6977 The default version of this function utilizes @code{ASM_GENERATE_INTERNAL_LABEL}.
6980 @defmac ASM_OUTPUT_DEBUG_LABEL (@var{stream}, @var{prefix}, @var{num})
6981 A C statement to output to the stdio stream @var{stream} a debug info
6982 label whose name is made from the string @var{prefix} and the number
6983 @var{num}. This is useful for VLIW targets, where debug info labels
6984 may need to be treated differently than branch target labels. On some
6985 systems, branch target labels must be at the beginning of instruction
6986 bundles, but debug info labels can occur in the middle of instruction
6989 If this macro is not defined, then @code{(*targetm.asm_out.internal_label)} will be
6993 @defmac ASM_GENERATE_INTERNAL_LABEL (@var{string}, @var{prefix}, @var{num})
6994 A C statement to store into the string @var{string} a label whose name
6995 is made from the string @var{prefix} and the number @var{num}.
6997 This string, when output subsequently by @code{assemble_name}, should
6998 produce the output that @code{(*targetm.asm_out.internal_label)} would produce
6999 with the same @var{prefix} and @var{num}.
7001 If the string begins with @samp{*}, then @code{assemble_name} will
7002 output the rest of the string unchanged. It is often convenient for
7003 @code{ASM_GENERATE_INTERNAL_LABEL} to use @samp{*} in this way. If the
7004 string doesn't start with @samp{*}, then @code{ASM_OUTPUT_LABELREF} gets
7005 to output the string, and may change it. (Of course,
7006 @code{ASM_OUTPUT_LABELREF} is also part of your machine description, so
7007 you should know what it does on your machine.)
7010 @defmac ASM_FORMAT_PRIVATE_NAME (@var{outvar}, @var{name}, @var{number})
7011 A C expression to assign to @var{outvar} (which is a variable of type
7012 @code{char *}) a newly allocated string made from the string
7013 @var{name} and the number @var{number}, with some suitable punctuation
7014 added. Use @code{alloca} to get space for the string.
7016 The string will be used as an argument to @code{ASM_OUTPUT_LABELREF} to
7017 produce an assembler label for an internal static variable whose name is
7018 @var{name}. Therefore, the string must be such as to result in valid
7019 assembler code. The argument @var{number} is different each time this
7020 macro is executed; it prevents conflicts between similarly-named
7021 internal static variables in different scopes.
7023 Ideally this string should not be a valid C identifier, to prevent any
7024 conflict with the user's own symbols. Most assemblers allow periods
7025 or percent signs in assembler symbols; putting at least one of these
7026 between the name and the number will suffice.
7028 If this macro is not defined, a default definition will be provided
7029 which is correct for most systems.
7032 @defmac ASM_OUTPUT_DEF (@var{stream}, @var{name}, @var{value})
7033 A C statement to output to the stdio stream @var{stream} assembler code
7034 which defines (equates) the symbol @var{name} to have the value @var{value}.
7037 If @code{SET_ASM_OP} is defined, a default definition is provided which is
7038 correct for most systems.
7041 @defmac ASM_OUTPUT_DEF_FROM_DECLS (@var{stream}, @var{decl_of_name}, @var{decl_of_value})
7042 A C statement to output to the stdio stream @var{stream} assembler code
7043 which defines (equates) the symbol whose tree node is @var{decl_of_name}
7044 to have the value of the tree node @var{decl_of_value}. This macro will
7045 be used in preference to @samp{ASM_OUTPUT_DEF} if it is defined and if
7046 the tree nodes are available.
7049 If @code{SET_ASM_OP} is defined, a default definition is provided which is
7050 correct for most systems.
7053 @defmac TARGET_DEFERRED_OUTPUT_DEFS (@var{decl_of_name}, @var{decl_of_value})
7054 A C statement that evaluates to true if the assembler code which defines
7055 (equates) the symbol whose tree node is @var{decl_of_name} to have the value
7056 of the tree node @var{decl_of_value} should be emitted near the end of the
7057 current compilation unit. The default is to not defer output of defines.
7058 This macro affects defines output by @samp{ASM_OUTPUT_DEF} and
7059 @samp{ASM_OUTPUT_DEF_FROM_DECLS}.
7062 @defmac ASM_OUTPUT_WEAK_ALIAS (@var{stream}, @var{name}, @var{value})
7063 A C statement to output to the stdio stream @var{stream} assembler code
7064 which defines (equates) the weak symbol @var{name} to have the value
7065 @var{value}. If @var{value} is @code{NULL}, it defines @var{name} as
7066 an undefined weak symbol.
7068 Define this macro if the target only supports weak aliases; define
7069 @code{ASM_OUTPUT_DEF} instead if possible.
7072 @defmac OBJC_GEN_METHOD_LABEL (@var{buf}, @var{is_inst}, @var{class_name}, @var{cat_name}, @var{sel_name})
7073 Define this macro to override the default assembler names used for
7074 Objective-C methods.
7076 The default name is a unique method number followed by the name of the
7077 class (e.g.@: @samp{_1_Foo}). For methods in categories, the name of
7078 the category is also included in the assembler name (e.g.@:
7081 These names are safe on most systems, but make debugging difficult since
7082 the method's selector is not present in the name. Therefore, particular
7083 systems define other ways of computing names.
7085 @var{buf} is an expression of type @code{char *} which gives you a
7086 buffer in which to store the name; its length is as long as
7087 @var{class_name}, @var{cat_name} and @var{sel_name} put together, plus
7088 50 characters extra.
7090 The argument @var{is_inst} specifies whether the method is an instance
7091 method or a class method; @var{class_name} is the name of the class;
7092 @var{cat_name} is the name of the category (or @code{NULL} if the method is not
7093 in a category); and @var{sel_name} is the name of the selector.
7095 On systems where the assembler can handle quoted names, you can use this
7096 macro to provide more human-readable names.
7099 @defmac ASM_DECLARE_CLASS_REFERENCE (@var{stream}, @var{name})
7100 A C statement (sans semicolon) to output to the stdio stream
7101 @var{stream} commands to declare that the label @var{name} is an
7102 Objective-C class reference. This is only needed for targets whose
7103 linkers have special support for NeXT-style runtimes.
7106 @defmac ASM_DECLARE_UNRESOLVED_REFERENCE (@var{stream}, @var{name})
7107 A C statement (sans semicolon) to output to the stdio stream
7108 @var{stream} commands to declare that the label @var{name} is an
7109 unresolved Objective-C class reference. This is only needed for targets
7110 whose linkers have special support for NeXT-style runtimes.
7113 @node Initialization
7114 @subsection How Initialization Functions Are Handled
7115 @cindex initialization routines
7116 @cindex termination routines
7117 @cindex constructors, output of
7118 @cindex destructors, output of
7120 The compiled code for certain languages includes @dfn{constructors}
7121 (also called @dfn{initialization routines})---functions to initialize
7122 data in the program when the program is started. These functions need
7123 to be called before the program is ``started''---that is to say, before
7124 @code{main} is called.
7126 Compiling some languages generates @dfn{destructors} (also called
7127 @dfn{termination routines}) that should be called when the program
7130 To make the initialization and termination functions work, the compiler
7131 must output something in the assembler code to cause those functions to
7132 be called at the appropriate time. When you port the compiler to a new
7133 system, you need to specify how to do this.
7135 There are two major ways that GCC currently supports the execution of
7136 initialization and termination functions. Each way has two variants.
7137 Much of the structure is common to all four variations.
7139 @findex __CTOR_LIST__
7140 @findex __DTOR_LIST__
7141 The linker must build two lists of these functions---a list of
7142 initialization functions, called @code{__CTOR_LIST__}, and a list of
7143 termination functions, called @code{__DTOR_LIST__}.
7145 Each list always begins with an ignored function pointer (which may hold
7146 0, @minus{}1, or a count of the function pointers after it, depending on
7147 the environment). This is followed by a series of zero or more function
7148 pointers to constructors (or destructors), followed by a function
7149 pointer containing zero.
7151 Depending on the operating system and its executable file format, either
7152 @file{crtstuff.c} or @file{libgcc2.c} traverses these lists at startup
7153 time and exit time. Constructors are called in reverse order of the
7154 list; destructors in forward order.
7156 The best way to handle static constructors works only for object file
7157 formats which provide arbitrarily-named sections. A section is set
7158 aside for a list of constructors, and another for a list of destructors.
7159 Traditionally these are called @samp{.ctors} and @samp{.dtors}. Each
7160 object file that defines an initialization function also puts a word in
7161 the constructor section to point to that function. The linker
7162 accumulates all these words into one contiguous @samp{.ctors} section.
7163 Termination functions are handled similarly.
7165 This method will be chosen as the default by @file{target-def.h} if
7166 @code{TARGET_ASM_NAMED_SECTION} is defined. A target that does not
7167 support arbitrary sections, but does support special designated
7168 constructor and destructor sections may define @code{CTORS_SECTION_ASM_OP}
7169 and @code{DTORS_SECTION_ASM_OP} to achieve the same effect.
7171 When arbitrary sections are available, there are two variants, depending
7172 upon how the code in @file{crtstuff.c} is called. On systems that
7173 support a @dfn{.init} section which is executed at program startup,
7174 parts of @file{crtstuff.c} are compiled into that section. The
7175 program is linked by the @command{gcc} driver like this:
7178 ld -o @var{output_file} crti.o crtbegin.o @dots{} -lgcc crtend.o crtn.o
7181 The prologue of a function (@code{__init}) appears in the @code{.init}
7182 section of @file{crti.o}; the epilogue appears in @file{crtn.o}. Likewise
7183 for the function @code{__fini} in the @dfn{.fini} section. Normally these
7184 files are provided by the operating system or by the GNU C library, but
7185 are provided by GCC for a few targets.
7187 The objects @file{crtbegin.o} and @file{crtend.o} are (for most targets)
7188 compiled from @file{crtstuff.c}. They contain, among other things, code
7189 fragments within the @code{.init} and @code{.fini} sections that branch
7190 to routines in the @code{.text} section. The linker will pull all parts
7191 of a section together, which results in a complete @code{__init} function
7192 that invokes the routines we need at startup.
7194 To use this variant, you must define the @code{INIT_SECTION_ASM_OP}
7197 If no init section is available, when GCC compiles any function called
7198 @code{main} (or more accurately, any function designated as a program
7199 entry point by the language front end calling @code{expand_main_function}),
7200 it inserts a procedure call to @code{__main} as the first executable code
7201 after the function prologue. The @code{__main} function is defined
7202 in @file{libgcc2.c} and runs the global constructors.
7204 In file formats that don't support arbitrary sections, there are again
7205 two variants. In the simplest variant, the GNU linker (GNU @code{ld})
7206 and an `a.out' format must be used. In this case,
7207 @code{TARGET_ASM_CONSTRUCTOR} is defined to produce a @code{.stabs}
7208 entry of type @samp{N_SETT}, referencing the name @code{__CTOR_LIST__},
7209 and with the address of the void function containing the initialization
7210 code as its value. The GNU linker recognizes this as a request to add
7211 the value to a @dfn{set}; the values are accumulated, and are eventually
7212 placed in the executable as a vector in the format described above, with
7213 a leading (ignored) count and a trailing zero element.
7214 @code{TARGET_ASM_DESTRUCTOR} is handled similarly. Since no init
7215 section is available, the absence of @code{INIT_SECTION_ASM_OP} causes
7216 the compilation of @code{main} to call @code{__main} as above, starting
7217 the initialization process.
7219 The last variant uses neither arbitrary sections nor the GNU linker.
7220 This is preferable when you want to do dynamic linking and when using
7221 file formats which the GNU linker does not support, such as `ECOFF'@. In
7222 this case, @code{TARGET_HAVE_CTORS_DTORS} is false, initialization and
7223 termination functions are recognized simply by their names. This requires
7224 an extra program in the linkage step, called @command{collect2}. This program
7225 pretends to be the linker, for use with GCC; it does its job by running
7226 the ordinary linker, but also arranges to include the vectors of
7227 initialization and termination functions. These functions are called
7228 via @code{__main} as described above. In order to use this method,
7229 @code{use_collect2} must be defined in the target in @file{config.gcc}.
7232 The following section describes the specific macros that control and
7233 customize the handling of initialization and termination functions.
7236 @node Macros for Initialization
7237 @subsection Macros Controlling Initialization Routines
7239 Here are the macros that control how the compiler handles initialization
7240 and termination functions:
7242 @defmac INIT_SECTION_ASM_OP
7243 If defined, a C string constant, including spacing, for the assembler
7244 operation to identify the following data as initialization code. If not
7245 defined, GCC will assume such a section does not exist. When you are
7246 using special sections for initialization and termination functions, this
7247 macro also controls how @file{crtstuff.c} and @file{libgcc2.c} arrange to
7248 run the initialization functions.
7251 @defmac HAS_INIT_SECTION
7252 If defined, @code{main} will not call @code{__main} as described above.
7253 This macro should be defined for systems that control start-up code
7254 on a symbol-by-symbol basis, such as OSF/1, and should not
7255 be defined explicitly for systems that support @code{INIT_SECTION_ASM_OP}.
7258 @defmac LD_INIT_SWITCH
7259 If defined, a C string constant for a switch that tells the linker that
7260 the following symbol is an initialization routine.
7263 @defmac LD_FINI_SWITCH
7264 If defined, a C string constant for a switch that tells the linker that
7265 the following symbol is a finalization routine.
7268 @defmac COLLECT_SHARED_INIT_FUNC (@var{stream}, @var{func})
7269 If defined, a C statement that will write a function that can be
7270 automatically called when a shared library is loaded. The function
7271 should call @var{func}, which takes no arguments. If not defined, and
7272 the object format requires an explicit initialization function, then a
7273 function called @code{_GLOBAL__DI} will be generated.
7275 This function and the following one are used by collect2 when linking a
7276 shared library that needs constructors or destructors, or has DWARF2
7277 exception tables embedded in the code.
7280 @defmac COLLECT_SHARED_FINI_FUNC (@var{stream}, @var{func})
7281 If defined, a C statement that will write a function that can be
7282 automatically called when a shared library is unloaded. The function
7283 should call @var{func}, which takes no arguments. If not defined, and
7284 the object format requires an explicit finalization function, then a
7285 function called @code{_GLOBAL__DD} will be generated.
7288 @defmac INVOKE__main
7289 If defined, @code{main} will call @code{__main} despite the presence of
7290 @code{INIT_SECTION_ASM_OP}. This macro should be defined for systems
7291 where the init section is not actually run automatically, but is still
7292 useful for collecting the lists of constructors and destructors.
7295 @defmac SUPPORTS_INIT_PRIORITY
7296 If nonzero, the C++ @code{init_priority} attribute is supported and the
7297 compiler should emit instructions to control the order of initialization
7298 of objects. If zero, the compiler will issue an error message upon
7299 encountering an @code{init_priority} attribute.
7302 @deftypefn {Target Hook} bool TARGET_HAVE_CTORS_DTORS
7303 This value is true if the target supports some ``native'' method of
7304 collecting constructors and destructors to be run at startup and exit.
7305 It is false if we must use @command{collect2}.
7308 @deftypefn {Target Hook} void TARGET_ASM_CONSTRUCTOR (rtx @var{symbol}, int @var{priority})
7309 If defined, a function that outputs assembler code to arrange to call
7310 the function referenced by @var{symbol} at initialization time.
7312 Assume that @var{symbol} is a @code{SYMBOL_REF} for a function taking
7313 no arguments and with no return value. If the target supports initialization
7314 priorities, @var{priority} is a value between 0 and @code{MAX_INIT_PRIORITY};
7315 otherwise it must be @code{DEFAULT_INIT_PRIORITY}.
7317 If this macro is not defined by the target, a suitable default will
7318 be chosen if (1) the target supports arbitrary section names, (2) the
7319 target defines @code{CTORS_SECTION_ASM_OP}, or (3) @code{USE_COLLECT2}
7323 @deftypefn {Target Hook} void TARGET_ASM_DESTRUCTOR (rtx @var{symbol}, int @var{priority})
7324 This is like @code{TARGET_ASM_CONSTRUCTOR} but used for termination
7325 functions rather than initialization functions.
7328 If @code{TARGET_HAVE_CTORS_DTORS} is true, the initialization routine
7329 generated for the generated object file will have static linkage.
7331 If your system uses @command{collect2} as the means of processing
7332 constructors, then that program normally uses @command{nm} to scan
7333 an object file for constructor functions to be called.
7335 On certain kinds of systems, you can define this macro to make
7336 @command{collect2} work faster (and, in some cases, make it work at all):
7338 @defmac OBJECT_FORMAT_COFF
7339 Define this macro if the system uses COFF (Common Object File Format)
7340 object files, so that @command{collect2} can assume this format and scan
7341 object files directly for dynamic constructor/destructor functions.
7343 This macro is effective only in a native compiler; @command{collect2} as
7344 part of a cross compiler always uses @command{nm} for the target machine.
7347 @defmac REAL_NM_FILE_NAME
7348 Define this macro as a C string constant containing the file name to use
7349 to execute @command{nm}. The default is to search the path normally for
7352 If your system supports shared libraries and has a program to list the
7353 dynamic dependencies of a given library or executable, you can define
7354 these macros to enable support for running initialization and
7355 termination functions in shared libraries:
7359 Define this macro to a C string constant containing the name of the program
7360 which lists dynamic dependencies, like @command{"ldd"} under SunOS 4.
7363 @defmac PARSE_LDD_OUTPUT (@var{ptr})
7364 Define this macro to be C code that extracts filenames from the output
7365 of the program denoted by @code{LDD_SUFFIX}. @var{ptr} is a variable
7366 of type @code{char *} that points to the beginning of a line of output
7367 from @code{LDD_SUFFIX}. If the line lists a dynamic dependency, the
7368 code must advance @var{ptr} to the beginning of the filename on that
7369 line. Otherwise, it must set @var{ptr} to @code{NULL}.
7372 @node Instruction Output
7373 @subsection Output of Assembler Instructions
7375 @c prevent bad page break with this line
7376 This describes assembler instruction output.
7378 @defmac REGISTER_NAMES
7379 A C initializer containing the assembler's names for the machine
7380 registers, each one as a C string constant. This is what translates
7381 register numbers in the compiler into assembler language.
7384 @defmac ADDITIONAL_REGISTER_NAMES
7385 If defined, a C initializer for an array of structures containing a name
7386 and a register number. This macro defines additional names for hard
7387 registers, thus allowing the @code{asm} option in declarations to refer
7388 to registers using alternate names.
7391 @defmac ASM_OUTPUT_OPCODE (@var{stream}, @var{ptr})
7392 Define this macro if you are using an unusual assembler that
7393 requires different names for the machine instructions.
7395 The definition is a C statement or statements which output an
7396 assembler instruction opcode to the stdio stream @var{stream}. The
7397 macro-operand @var{ptr} is a variable of type @code{char *} which
7398 points to the opcode name in its ``internal'' form---the form that is
7399 written in the machine description. The definition should output the
7400 opcode name to @var{stream}, performing any translation you desire, and
7401 increment the variable @var{ptr} to point at the end of the opcode
7402 so that it will not be output twice.
7404 In fact, your macro definition may process less than the entire opcode
7405 name, or more than the opcode name; but if you want to process text
7406 that includes @samp{%}-sequences to substitute operands, you must take
7407 care of the substitution yourself. Just be sure to increment
7408 @var{ptr} over whatever text should not be output normally.
7410 @findex recog_data.operand
7411 If you need to look at the operand values, they can be found as the
7412 elements of @code{recog_data.operand}.
7414 If the macro definition does nothing, the instruction is output
7418 @defmac FINAL_PRESCAN_INSN (@var{insn}, @var{opvec}, @var{noperands})
7419 If defined, a C statement to be executed just prior to the output of
7420 assembler code for @var{insn}, to modify the extracted operands so
7421 they will be output differently.
7423 Here the argument @var{opvec} is the vector containing the operands
7424 extracted from @var{insn}, and @var{noperands} is the number of
7425 elements of the vector which contain meaningful data for this insn.
7426 The contents of this vector are what will be used to convert the insn
7427 template into assembler code, so you can change the assembler output
7428 by changing the contents of the vector.
7430 This macro is useful when various assembler syntaxes share a single
7431 file of instruction patterns; by defining this macro differently, you
7432 can cause a large class of instructions to be output differently (such
7433 as with rearranged operands). Naturally, variations in assembler
7434 syntax affecting individual insn patterns ought to be handled by
7435 writing conditional output routines in those patterns.
7437 If this macro is not defined, it is equivalent to a null statement.
7440 @defmac PRINT_OPERAND (@var{stream}, @var{x}, @var{code})
7441 A C compound statement to output to stdio stream @var{stream} the
7442 assembler syntax for an instruction operand @var{x}. @var{x} is an
7445 @var{code} is a value that can be used to specify one of several ways
7446 of printing the operand. It is used when identical operands must be
7447 printed differently depending on the context. @var{code} comes from
7448 the @samp{%} specification that was used to request printing of the
7449 operand. If the specification was just @samp{%@var{digit}} then
7450 @var{code} is 0; if the specification was @samp{%@var{ltr}
7451 @var{digit}} then @var{code} is the ASCII code for @var{ltr}.
7454 If @var{x} is a register, this macro should print the register's name.
7455 The names can be found in an array @code{reg_names} whose type is
7456 @code{char *[]}. @code{reg_names} is initialized from
7457 @code{REGISTER_NAMES}.
7459 When the machine description has a specification @samp{%@var{punct}}
7460 (a @samp{%} followed by a punctuation character), this macro is called
7461 with a null pointer for @var{x} and the punctuation character for
7465 @defmac PRINT_OPERAND_PUNCT_VALID_P (@var{code})
7466 A C expression which evaluates to true if @var{code} is a valid
7467 punctuation character for use in the @code{PRINT_OPERAND} macro. If
7468 @code{PRINT_OPERAND_PUNCT_VALID_P} is not defined, it means that no
7469 punctuation characters (except for the standard one, @samp{%}) are used
7473 @defmac PRINT_OPERAND_ADDRESS (@var{stream}, @var{x})
7474 A C compound statement to output to stdio stream @var{stream} the
7475 assembler syntax for an instruction operand that is a memory reference
7476 whose address is @var{x}. @var{x} is an RTL expression.
7478 @cindex @code{TARGET_ENCODE_SECTION_INFO} usage
7479 On some machines, the syntax for a symbolic address depends on the
7480 section that the address refers to. On these machines, define the hook
7481 @code{TARGET_ENCODE_SECTION_INFO} to store the information into the
7482 @code{symbol_ref}, and then check for it here. @xref{Assembler
7486 @findex dbr_sequence_length
7487 @defmac DBR_OUTPUT_SEQEND (@var{file})
7488 A C statement, to be executed after all slot-filler instructions have
7489 been output. If necessary, call @code{dbr_sequence_length} to
7490 determine the number of slots filled in a sequence (zero if not
7491 currently outputting a sequence), to decide how many no-ops to output,
7494 Don't define this macro if it has nothing to do, but it is helpful in
7495 reading assembly output if the extent of the delay sequence is made
7496 explicit (e.g.@: with white space).
7499 @findex final_sequence
7500 Note that output routines for instructions with delay slots must be
7501 prepared to deal with not being output as part of a sequence
7502 (i.e.@: when the scheduling pass is not run, or when no slot fillers could be
7503 found.) The variable @code{final_sequence} is null when not
7504 processing a sequence, otherwise it contains the @code{sequence} rtx
7508 @defmac REGISTER_PREFIX
7509 @defmacx LOCAL_LABEL_PREFIX
7510 @defmacx USER_LABEL_PREFIX
7511 @defmacx IMMEDIATE_PREFIX
7512 If defined, C string expressions to be used for the @samp{%R}, @samp{%L},
7513 @samp{%U}, and @samp{%I} options of @code{asm_fprintf} (see
7514 @file{final.c}). These are useful when a single @file{md} file must
7515 support multiple assembler formats. In that case, the various @file{tm.h}
7516 files can define these macros differently.
7519 @defmac ASM_FPRINTF_EXTENSIONS (@var{file}, @var{argptr}, @var{format})
7520 If defined this macro should expand to a series of @code{case}
7521 statements which will be parsed inside the @code{switch} statement of
7522 the @code{asm_fprintf} function. This allows targets to define extra
7523 printf formats which may useful when generating their assembler
7524 statements. Note that uppercase letters are reserved for future
7525 generic extensions to asm_fprintf, and so are not available to target
7526 specific code. The output file is given by the parameter @var{file}.
7527 The varargs input pointer is @var{argptr} and the rest of the format
7528 string, starting the character after the one that is being switched
7529 upon, is pointed to by @var{format}.
7532 @defmac ASSEMBLER_DIALECT
7533 If your target supports multiple dialects of assembler language (such as
7534 different opcodes), define this macro as a C expression that gives the
7535 numeric index of the assembler language dialect to use, with zero as the
7538 If this macro is defined, you may use constructs of the form
7540 @samp{@{option0|option1|option2@dots{}@}}
7543 in the output templates of patterns (@pxref{Output Template}) or in the
7544 first argument of @code{asm_fprintf}. This construct outputs
7545 @samp{option0}, @samp{option1}, @samp{option2}, etc., if the value of
7546 @code{ASSEMBLER_DIALECT} is zero, one, two, etc. Any special characters
7547 within these strings retain their usual meaning. If there are fewer
7548 alternatives within the braces than the value of
7549 @code{ASSEMBLER_DIALECT}, the construct outputs nothing.
7551 If you do not define this macro, the characters @samp{@{}, @samp{|} and
7552 @samp{@}} do not have any special meaning when used in templates or
7553 operands to @code{asm_fprintf}.
7555 Define the macros @code{REGISTER_PREFIX}, @code{LOCAL_LABEL_PREFIX},
7556 @code{USER_LABEL_PREFIX} and @code{IMMEDIATE_PREFIX} if you can express
7557 the variations in assembler language syntax with that mechanism. Define
7558 @code{ASSEMBLER_DIALECT} and use the @samp{@{option0|option1@}} syntax
7559 if the syntax variant are larger and involve such things as different
7560 opcodes or operand order.
7563 @defmac ASM_OUTPUT_REG_PUSH (@var{stream}, @var{regno})
7564 A C expression to output to @var{stream} some assembler code
7565 which will push hard register number @var{regno} onto the stack.
7566 The code need not be optimal, since this macro is used only when
7570 @defmac ASM_OUTPUT_REG_POP (@var{stream}, @var{regno})
7571 A C expression to output to @var{stream} some assembler code
7572 which will pop hard register number @var{regno} off of the stack.
7573 The code need not be optimal, since this macro is used only when
7577 @node Dispatch Tables
7578 @subsection Output of Dispatch Tables
7580 @c prevent bad page break with this line
7581 This concerns dispatch tables.
7583 @cindex dispatch table
7584 @defmac ASM_OUTPUT_ADDR_DIFF_ELT (@var{stream}, @var{body}, @var{value}, @var{rel})
7585 A C statement to output to the stdio stream @var{stream} an assembler
7586 pseudo-instruction to generate a difference between two labels.
7587 @var{value} and @var{rel} are the numbers of two internal labels. The
7588 definitions of these labels are output using
7589 @code{(*targetm.asm_out.internal_label)}, and they must be printed in the same
7590 way here. For example,
7593 fprintf (@var{stream}, "\t.word L%d-L%d\n",
7594 @var{value}, @var{rel})
7597 You must provide this macro on machines where the addresses in a
7598 dispatch table are relative to the table's own address. If defined, GCC
7599 will also use this macro on all machines when producing PIC@.
7600 @var{body} is the body of the @code{ADDR_DIFF_VEC}; it is provided so that the
7601 mode and flags can be read.
7604 @defmac ASM_OUTPUT_ADDR_VEC_ELT (@var{stream}, @var{value})
7605 This macro should be provided on machines where the addresses
7606 in a dispatch table are absolute.
7608 The definition should be a C statement to output to the stdio stream
7609 @var{stream} an assembler pseudo-instruction to generate a reference to
7610 a label. @var{value} is the number of an internal label whose
7611 definition is output using @code{(*targetm.asm_out.internal_label)}.
7615 fprintf (@var{stream}, "\t.word L%d\n", @var{value})
7619 @defmac ASM_OUTPUT_CASE_LABEL (@var{stream}, @var{prefix}, @var{num}, @var{table})
7620 Define this if the label before a jump-table needs to be output
7621 specially. The first three arguments are the same as for
7622 @code{(*targetm.asm_out.internal_label)}; the fourth argument is the
7623 jump-table which follows (a @code{jump_insn} containing an
7624 @code{addr_vec} or @code{addr_diff_vec}).
7626 This feature is used on system V to output a @code{swbeg} statement
7629 If this macro is not defined, these labels are output with
7630 @code{(*targetm.asm_out.internal_label)}.
7633 @defmac ASM_OUTPUT_CASE_END (@var{stream}, @var{num}, @var{table})
7634 Define this if something special must be output at the end of a
7635 jump-table. The definition should be a C statement to be executed
7636 after the assembler code for the table is written. It should write
7637 the appropriate code to stdio stream @var{stream}. The argument
7638 @var{table} is the jump-table insn, and @var{num} is the label-number
7639 of the preceding label.
7641 If this macro is not defined, nothing special is output at the end of
7645 @deftypefn {Target Hook} void TARGET_ASM_EMIT_UNWIND_LABEL (@var{stream}, @var{decl}, @var{for_eh}, @var{empty})
7646 This target hook emits a label at the beginning of each FDE@. It
7647 should be defined on targets where FDEs need special labels, and it
7648 should write the appropriate label, for the FDE associated with the
7649 function declaration @var{decl}, to the stdio stream @var{stream}.
7650 The third argument, @var{for_eh}, is a boolean: true if this is for an
7651 exception table. The fourth argument, @var{empty}, is a boolean:
7652 true if this is a placeholder label for an omitted FDE@.
7654 The default is that FDEs are not given nonlocal labels.
7657 @deftypefn {Taget Hook} void TARGET_UNWIND_EMIT (FILE * @var{stream}, rtx @var{insn})
7658 This target hook emits and assembly directives required to unwind the
7659 given instruction. This is only used when TARGET_UNWIND_INFO is set.
7662 @node Exception Region Output
7663 @subsection Assembler Commands for Exception Regions
7665 @c prevent bad page break with this line
7667 This describes commands marking the start and the end of an exception
7670 @defmac EH_FRAME_SECTION_NAME
7671 If defined, a C string constant for the name of the section containing
7672 exception handling frame unwind information. If not defined, GCC will
7673 provide a default definition if the target supports named sections.
7674 @file{crtstuff.c} uses this macro to switch to the appropriate section.
7676 You should define this symbol if your target supports DWARF 2 frame
7677 unwind information and the default definition does not work.
7680 @defmac EH_FRAME_IN_DATA_SECTION
7681 If defined, DWARF 2 frame unwind information will be placed in the
7682 data section even though the target supports named sections. This
7683 might be necessary, for instance, if the system linker does garbage
7684 collection and sections cannot be marked as not to be collected.
7686 Do not define this macro unless @code{TARGET_ASM_NAMED_SECTION} is
7690 @defmac EH_TABLES_CAN_BE_READ_ONLY
7691 Define this macro to 1 if your target is such that no frame unwind
7692 information encoding used with non-PIC code will ever require a
7693 runtime relocation, but the linker may not support merging read-only
7694 and read-write sections into a single read-write section.
7697 @defmac MASK_RETURN_ADDR
7698 An rtx used to mask the return address found via @code{RETURN_ADDR_RTX}, so
7699 that it does not contain any extraneous set bits in it.
7702 @defmac DWARF2_UNWIND_INFO
7703 Define this macro to 0 if your target supports DWARF 2 frame unwind
7704 information, but it does not yet work with exception handling.
7705 Otherwise, if your target supports this information (if it defines
7706 @samp{INCOMING_RETURN_ADDR_RTX} and either @samp{UNALIGNED_INT_ASM_OP}
7707 or @samp{OBJECT_FORMAT_ELF}), GCC will provide a default definition of
7710 If @code{TARGET_UNWIND_INFO} is defined, the target specific unwinder
7711 will be used in all cases. Defining this macro will enable the generation
7712 of DWARF 2 frame debugging information.
7714 If @code{TARGET_UNWIND_INFO} is not defined, and this macro is defined to 1,
7715 the DWARF 2 unwinder will be the default exception handling mechanism;
7716 otherwise, @code{setjmp}/@code{longjmp} will be used by default.
7719 @defmac TARGET_UNWIND_INFO
7720 Define this macro if your target has ABI specified unwind tables. Usually
7721 these will be output by @code{TARGET_UNWIND_EMIT}.
7724 @deftypevar {Target Hook} bool TARGET_UNWID_TABLES_DEFAULT
7725 This variable should be set to @code{true} if the target ABI requires unwinding
7726 tables even when exceptions are not used.
7729 @defmac MUST_USE_SJLJ_EXCEPTIONS
7730 This macro need only be defined if @code{DWARF2_UNWIND_INFO} is
7731 runtime-variable. In that case, @file{except.h} cannot correctly
7732 determine the corresponding definition of
7733 @code{MUST_USE_SJLJ_EXCEPTIONS}, so the target must provide it directly.
7736 @defmac DWARF_CIE_DATA_ALIGNMENT
7737 This macro need only be defined if the target might save registers in the
7738 function prologue at an offset to the stack pointer that is not aligned to
7739 @code{UNITS_PER_WORD}. The definition should be the negative minimum
7740 alignment if @code{STACK_GROWS_DOWNWARD} is defined, and the positive
7741 minimum alignment otherwise. @xref{SDB and DWARF}. Only applicable if
7742 the target supports DWARF 2 frame unwind information.
7745 @deftypefn {Target Hook} void TARGET_ASM_EXCEPTION_SECTION ()
7746 If defined, a function that switches to the section in which the main
7747 exception table is to be placed (@pxref{Sections}). The default is a
7748 function that switches to a section named @code{.gcc_except_table} on
7749 machines that support named sections via
7750 @code{TARGET_ASM_NAMED_SECTION}, otherwise if @option{-fpic} or
7751 @option{-fPIC} is in effect, the @code{data_section}, otherwise the
7752 @code{readonly_data_section}.
7755 @deftypefn {Target Hook} void TARGET_ASM_EH_FRAME_SECTION ()
7756 If defined, a function that switches to the section in which the DWARF 2
7757 frame unwind information to be placed (@pxref{Sections}). The default
7758 is a function that outputs a standard GAS section directive, if
7759 @code{EH_FRAME_SECTION_NAME} is defined, or else a data section
7760 directive followed by a synthetic label.
7763 @deftypevar {Target Hook} bool TARGET_TERMINATE_DW2_EH_FRAME_INFO
7764 Contains the value true if the target should add a zero word onto the
7765 end of a Dwarf-2 frame info section when used for exception handling.
7766 Default value is false if @code{EH_FRAME_SECTION_NAME} is defined, and
7770 @deftypefn {Target Hook} rtx TARGET_DWARF_REGISTER_SPAN (rtx @var{reg})
7771 Given a register, this hook should return a parallel of registers to
7772 represent where to find the register pieces. Define this hook if the
7773 register and its mode are represented in Dwarf in non-contiguous
7774 locations, or if the register should be represented in more than one
7775 register in Dwarf. Otherwise, this hook should return @code{NULL_RTX}.
7776 If not defined, the default is to return @code{NULL_RTX}.
7779 @deftypefn {Target Hook} bool TARGET_ASM_TTYPE (rtx @var{sym})
7780 This hook is used to output a reference from a frame unwinding table to
7781 the type_info object identified by @var{sym}. It should return @code{true}
7782 if the reference was output. Returning @code{false} will cause the
7783 reference to be output using the normal Dwarf2 routines.
7786 @deftypefn {Target Hook} bool TARGET_ARM_EABI_UNWINDER
7787 This hook should be set to @code{true} on targets that use an ARM EABI
7788 based unwinding library, and @code{false} on other targets. This effects
7789 the format of unwinding tables, and how the unwinder in entered after
7790 running a cleanup. The default is @code{false}.
7793 @node Alignment Output
7794 @subsection Assembler Commands for Alignment
7796 @c prevent bad page break with this line
7797 This describes commands for alignment.
7799 @defmac JUMP_ALIGN (@var{label})
7800 The alignment (log base 2) to put in front of @var{label}, which is
7801 a common destination of jumps and has no fallthru incoming edge.
7803 This macro need not be defined if you don't want any special alignment
7804 to be done at such a time. Most machine descriptions do not currently
7807 Unless it's necessary to inspect the @var{label} parameter, it is better
7808 to set the variable @var{align_jumps} in the target's
7809 @code{OVERRIDE_OPTIONS}. Otherwise, you should try to honor the user's
7810 selection in @var{align_jumps} in a @code{JUMP_ALIGN} implementation.
7813 @defmac LABEL_ALIGN_AFTER_BARRIER (@var{label})
7814 The alignment (log base 2) to put in front of @var{label}, which follows
7817 This macro need not be defined if you don't want any special alignment
7818 to be done at such a time. Most machine descriptions do not currently
7822 @defmac LABEL_ALIGN_AFTER_BARRIER_MAX_SKIP
7823 The maximum number of bytes to skip when applying
7824 @code{LABEL_ALIGN_AFTER_BARRIER}. This works only if
7825 @code{ASM_OUTPUT_MAX_SKIP_ALIGN} is defined.
7828 @defmac LOOP_ALIGN (@var{label})
7829 The alignment (log base 2) to put in front of @var{label}, which follows
7830 a @code{NOTE_INSN_LOOP_BEG} note.
7832 This macro need not be defined if you don't want any special alignment
7833 to be done at such a time. Most machine descriptions do not currently
7836 Unless it's necessary to inspect the @var{label} parameter, it is better
7837 to set the variable @code{align_loops} in the target's
7838 @code{OVERRIDE_OPTIONS}. Otherwise, you should try to honor the user's
7839 selection in @code{align_loops} in a @code{LOOP_ALIGN} implementation.
7842 @defmac LOOP_ALIGN_MAX_SKIP
7843 The maximum number of bytes to skip when applying @code{LOOP_ALIGN}.
7844 This works only if @code{ASM_OUTPUT_MAX_SKIP_ALIGN} is defined.
7847 @defmac LABEL_ALIGN (@var{label})
7848 The alignment (log base 2) to put in front of @var{label}.
7849 If @code{LABEL_ALIGN_AFTER_BARRIER} / @code{LOOP_ALIGN} specify a different alignment,
7850 the maximum of the specified values is used.
7852 Unless it's necessary to inspect the @var{label} parameter, it is better
7853 to set the variable @code{align_labels} in the target's
7854 @code{OVERRIDE_OPTIONS}. Otherwise, you should try to honor the user's
7855 selection in @code{align_labels} in a @code{LABEL_ALIGN} implementation.
7858 @defmac LABEL_ALIGN_MAX_SKIP
7859 The maximum number of bytes to skip when applying @code{LABEL_ALIGN}.
7860 This works only if @code{ASM_OUTPUT_MAX_SKIP_ALIGN} is defined.
7863 @defmac ASM_OUTPUT_SKIP (@var{stream}, @var{nbytes})
7864 A C statement to output to the stdio stream @var{stream} an assembler
7865 instruction to advance the location counter by @var{nbytes} bytes.
7866 Those bytes should be zero when loaded. @var{nbytes} will be a C
7867 expression of type @code{int}.
7870 @defmac ASM_NO_SKIP_IN_TEXT
7871 Define this macro if @code{ASM_OUTPUT_SKIP} should not be used in the
7872 text section because it fails to put zeros in the bytes that are skipped.
7873 This is true on many Unix systems, where the pseudo--op to skip bytes
7874 produces no-op instructions rather than zeros when used in the text
7878 @defmac ASM_OUTPUT_ALIGN (@var{stream}, @var{power})
7879 A C statement to output to the stdio stream @var{stream} an assembler
7880 command to advance the location counter to a multiple of 2 to the
7881 @var{power} bytes. @var{power} will be a C expression of type @code{int}.
7884 @defmac ASM_OUTPUT_ALIGN_WITH_NOP (@var{stream}, @var{power})
7885 Like @code{ASM_OUTPUT_ALIGN}, except that the ``nop'' instruction is used
7886 for padding, if necessary.
7889 @defmac ASM_OUTPUT_MAX_SKIP_ALIGN (@var{stream}, @var{power}, @var{max_skip})
7890 A C statement to output to the stdio stream @var{stream} an assembler
7891 command to advance the location counter to a multiple of 2 to the
7892 @var{power} bytes, but only if @var{max_skip} or fewer bytes are needed to
7893 satisfy the alignment request. @var{power} and @var{max_skip} will be
7894 a C expression of type @code{int}.
7898 @node Debugging Info
7899 @section Controlling Debugging Information Format
7901 @c prevent bad page break with this line
7902 This describes how to specify debugging information.
7905 * All Debuggers:: Macros that affect all debugging formats uniformly.
7906 * DBX Options:: Macros enabling specific options in DBX format.
7907 * DBX Hooks:: Hook macros for varying DBX format.
7908 * File Names and DBX:: Macros controlling output of file names in DBX format.
7909 * SDB and DWARF:: Macros for SDB (COFF) and DWARF formats.
7910 * VMS Debug:: Macros for VMS debug format.
7914 @subsection Macros Affecting All Debugging Formats
7916 @c prevent bad page break with this line
7917 These macros affect all debugging formats.
7919 @defmac DBX_REGISTER_NUMBER (@var{regno})
7920 A C expression that returns the DBX register number for the compiler
7921 register number @var{regno}. In the default macro provided, the value
7922 of this expression will be @var{regno} itself. But sometimes there are
7923 some registers that the compiler knows about and DBX does not, or vice
7924 versa. In such cases, some register may need to have one number in the
7925 compiler and another for DBX@.
7927 If two registers have consecutive numbers inside GCC, and they can be
7928 used as a pair to hold a multiword value, then they @emph{must} have
7929 consecutive numbers after renumbering with @code{DBX_REGISTER_NUMBER}.
7930 Otherwise, debuggers will be unable to access such a pair, because they
7931 expect register pairs to be consecutive in their own numbering scheme.
7933 If you find yourself defining @code{DBX_REGISTER_NUMBER} in way that
7934 does not preserve register pairs, then what you must do instead is
7935 redefine the actual register numbering scheme.
7938 @defmac DEBUGGER_AUTO_OFFSET (@var{x})
7939 A C expression that returns the integer offset value for an automatic
7940 variable having address @var{x} (an RTL expression). The default
7941 computation assumes that @var{x} is based on the frame-pointer and
7942 gives the offset from the frame-pointer. This is required for targets
7943 that produce debugging output for DBX or COFF-style debugging output
7944 for SDB and allow the frame-pointer to be eliminated when the
7945 @option{-g} options is used.
7948 @defmac DEBUGGER_ARG_OFFSET (@var{offset}, @var{x})
7949 A C expression that returns the integer offset value for an argument
7950 having address @var{x} (an RTL expression). The nominal offset is
7954 @defmac PREFERRED_DEBUGGING_TYPE
7955 A C expression that returns the type of debugging output GCC should
7956 produce when the user specifies just @option{-g}. Define
7957 this if you have arranged for GCC to support more than one format of
7958 debugging output. Currently, the allowable values are @code{DBX_DEBUG},
7959 @code{SDB_DEBUG}, @code{DWARF_DEBUG}, @code{DWARF2_DEBUG},
7960 @code{XCOFF_DEBUG}, @code{VMS_DEBUG}, and @code{VMS_AND_DWARF2_DEBUG}.
7962 When the user specifies @option{-ggdb}, GCC normally also uses the
7963 value of this macro to select the debugging output format, but with two
7964 exceptions. If @code{DWARF2_DEBUGGING_INFO} is defined, GCC uses the
7965 value @code{DWARF2_DEBUG}. Otherwise, if @code{DBX_DEBUGGING_INFO} is
7966 defined, GCC uses @code{DBX_DEBUG}.
7968 The value of this macro only affects the default debugging output; the
7969 user can always get a specific type of output by using @option{-gstabs},
7970 @option{-gcoff}, @option{-gdwarf-2}, @option{-gxcoff}, or @option{-gvms}.
7974 @subsection Specific Options for DBX Output
7976 @c prevent bad page break with this line
7977 These are specific options for DBX output.
7979 @defmac DBX_DEBUGGING_INFO
7980 Define this macro if GCC should produce debugging output for DBX
7981 in response to the @option{-g} option.
7984 @defmac XCOFF_DEBUGGING_INFO
7985 Define this macro if GCC should produce XCOFF format debugging output
7986 in response to the @option{-g} option. This is a variant of DBX format.
7989 @defmac DEFAULT_GDB_EXTENSIONS
7990 Define this macro to control whether GCC should by default generate
7991 GDB's extended version of DBX debugging information (assuming DBX-format
7992 debugging information is enabled at all). If you don't define the
7993 macro, the default is 1: always generate the extended information
7994 if there is any occasion to.
7997 @defmac DEBUG_SYMS_TEXT
7998 Define this macro if all @code{.stabs} commands should be output while
7999 in the text section.
8002 @defmac ASM_STABS_OP
8003 A C string constant, including spacing, naming the assembler pseudo op to
8004 use instead of @code{"\t.stabs\t"} to define an ordinary debugging symbol.
8005 If you don't define this macro, @code{"\t.stabs\t"} is used. This macro
8006 applies only to DBX debugging information format.
8009 @defmac ASM_STABD_OP
8010 A C string constant, including spacing, naming the assembler pseudo op to
8011 use instead of @code{"\t.stabd\t"} to define a debugging symbol whose
8012 value is the current location. If you don't define this macro,
8013 @code{"\t.stabd\t"} is used. This macro applies only to DBX debugging
8017 @defmac ASM_STABN_OP
8018 A C string constant, including spacing, naming the assembler pseudo op to
8019 use instead of @code{"\t.stabn\t"} to define a debugging symbol with no
8020 name. If you don't define this macro, @code{"\t.stabn\t"} is used. This
8021 macro applies only to DBX debugging information format.
8024 @defmac DBX_NO_XREFS
8025 Define this macro if DBX on your system does not support the construct
8026 @samp{xs@var{tagname}}. On some systems, this construct is used to
8027 describe a forward reference to a structure named @var{tagname}.
8028 On other systems, this construct is not supported at all.
8031 @defmac DBX_CONTIN_LENGTH
8032 A symbol name in DBX-format debugging information is normally
8033 continued (split into two separate @code{.stabs} directives) when it
8034 exceeds a certain length (by default, 80 characters). On some
8035 operating systems, DBX requires this splitting; on others, splitting
8036 must not be done. You can inhibit splitting by defining this macro
8037 with the value zero. You can override the default splitting-length by
8038 defining this macro as an expression for the length you desire.
8041 @defmac DBX_CONTIN_CHAR
8042 Normally continuation is indicated by adding a @samp{\} character to
8043 the end of a @code{.stabs} string when a continuation follows. To use
8044 a different character instead, define this macro as a character
8045 constant for the character you want to use. Do not define this macro
8046 if backslash is correct for your system.
8049 @defmac DBX_STATIC_STAB_DATA_SECTION
8050 Define this macro if it is necessary to go to the data section before
8051 outputting the @samp{.stabs} pseudo-op for a non-global static
8055 @defmac DBX_TYPE_DECL_STABS_CODE
8056 The value to use in the ``code'' field of the @code{.stabs} directive
8057 for a typedef. The default is @code{N_LSYM}.
8060 @defmac DBX_STATIC_CONST_VAR_CODE
8061 The value to use in the ``code'' field of the @code{.stabs} directive
8062 for a static variable located in the text section. DBX format does not
8063 provide any ``right'' way to do this. The default is @code{N_FUN}.
8066 @defmac DBX_REGPARM_STABS_CODE
8067 The value to use in the ``code'' field of the @code{.stabs} directive
8068 for a parameter passed in registers. DBX format does not provide any
8069 ``right'' way to do this. The default is @code{N_RSYM}.
8072 @defmac DBX_REGPARM_STABS_LETTER
8073 The letter to use in DBX symbol data to identify a symbol as a parameter
8074 passed in registers. DBX format does not customarily provide any way to
8075 do this. The default is @code{'P'}.
8078 @defmac DBX_FUNCTION_FIRST
8079 Define this macro if the DBX information for a function and its
8080 arguments should precede the assembler code for the function. Normally,
8081 in DBX format, the debugging information entirely follows the assembler
8085 @defmac DBX_BLOCKS_FUNCTION_RELATIVE
8086 Define this macro, with value 1, if the value of a symbol describing
8087 the scope of a block (@code{N_LBRAC} or @code{N_RBRAC}) should be
8088 relative to the start of the enclosing function. Normally, GCC uses
8089 an absolute address.
8092 @defmac DBX_LINES_FUNCTION_RELATIVE
8093 Define this macro, with value 1, if the value of a symbol indicating
8094 the current line number (@code{N_SLINE}) should be relative to the
8095 start of the enclosing function. Normally, GCC uses an absolute address.
8098 @defmac DBX_USE_BINCL
8099 Define this macro if GCC should generate @code{N_BINCL} and
8100 @code{N_EINCL} stabs for included header files, as on Sun systems. This
8101 macro also directs GCC to output a type number as a pair of a file
8102 number and a type number within the file. Normally, GCC does not
8103 generate @code{N_BINCL} or @code{N_EINCL} stabs, and it outputs a single
8104 number for a type number.
8108 @subsection Open-Ended Hooks for DBX Format
8110 @c prevent bad page break with this line
8111 These are hooks for DBX format.
8113 @defmac DBX_OUTPUT_LBRAC (@var{stream}, @var{name})
8114 Define this macro to say how to output to @var{stream} the debugging
8115 information for the start of a scope level for variable names. The
8116 argument @var{name} is the name of an assembler symbol (for use with
8117 @code{assemble_name}) whose value is the address where the scope begins.
8120 @defmac DBX_OUTPUT_RBRAC (@var{stream}, @var{name})
8121 Like @code{DBX_OUTPUT_LBRAC}, but for the end of a scope level.
8124 @defmac DBX_OUTPUT_NFUN (@var{stream}, @var{lscope_label}, @var{decl})
8125 Define this macro if the target machine requires special handling to
8126 output an @code{N_FUN} entry for the function @var{decl}.
8129 @defmac DBX_OUTPUT_SOURCE_LINE (@var{stream}, @var{line}, @var{counter})
8130 A C statement to output DBX debugging information before code for line
8131 number @var{line} of the current source file to the stdio stream
8132 @var{stream}. @var{counter} is the number of time the macro was
8133 invoked, including the current invocation; it is intended to generate
8134 unique labels in the assembly output.
8136 This macro should not be defined if the default output is correct, or
8137 if it can be made correct by defining @code{DBX_LINES_FUNCTION_RELATIVE}.
8140 @defmac NO_DBX_FUNCTION_END
8141 Some stabs encapsulation formats (in particular ECOFF), cannot handle the
8142 @code{.stabs "",N_FUN,,0,0,Lscope-function-1} gdb dbx extension construct.
8143 On those machines, define this macro to turn this feature off without
8144 disturbing the rest of the gdb extensions.
8147 @defmac NO_DBX_BNSYM_ENSYM
8148 Some assemblers cannot handle the @code{.stabd BNSYM/ENSYM,0,0} gdb dbx
8149 extension construct. On those machines, define this macro to turn this
8150 feature off without disturbing the rest of the gdb extensions.
8153 @node File Names and DBX
8154 @subsection File Names in DBX Format
8156 @c prevent bad page break with this line
8157 This describes file names in DBX format.
8159 @defmac DBX_OUTPUT_MAIN_SOURCE_FILENAME (@var{stream}, @var{name})
8160 A C statement to output DBX debugging information to the stdio stream
8161 @var{stream}, which indicates that file @var{name} is the main source
8162 file---the file specified as the input file for compilation.
8163 This macro is called only once, at the beginning of compilation.
8165 This macro need not be defined if the standard form of output
8166 for DBX debugging information is appropriate.
8168 It may be necessary to refer to a label equal to the beginning of the
8169 text section. You can use @samp{assemble_name (stream, ltext_label_name)}
8170 to do so. If you do this, you must also set the variable
8171 @var{used_ltext_label_name} to @code{true}.
8174 @defmac NO_DBX_MAIN_SOURCE_DIRECTORY
8175 Define this macro, with value 1, if GCC should not emit an indication
8176 of the current directory for compilation and current source language at
8177 the beginning of the file.
8180 @defmac NO_DBX_GCC_MARKER
8181 Define this macro, with value 1, if GCC should not emit an indication
8182 that this object file was compiled by GCC@. The default is to emit
8183 an @code{N_OPT} stab at the beginning of every source file, with
8184 @samp{gcc2_compiled.} for the string and value 0.
8187 @defmac DBX_OUTPUT_MAIN_SOURCE_FILE_END (@var{stream}, @var{name})
8188 A C statement to output DBX debugging information at the end of
8189 compilation of the main source file @var{name}. Output should be
8190 written to the stdio stream @var{stream}.
8192 If you don't define this macro, nothing special is output at the end
8193 of compilation, which is correct for most machines.
8196 @defmac DBX_OUTPUT_NULL_N_SO_AT_MAIN_SOURCE_FILE_END
8197 Define this macro @emph{instead of} defining
8198 @code{DBX_OUTPUT_MAIN_SOURCE_FILE_END}, if what needs to be output at
8199 the end of compilation is a @code{N_SO} stab with an empty string,
8200 whose value is the highest absolute text address in the file.
8205 @subsection Macros for SDB and DWARF Output
8207 @c prevent bad page break with this line
8208 Here are macros for SDB and DWARF output.
8210 @defmac SDB_DEBUGGING_INFO
8211 Define this macro if GCC should produce COFF-style debugging output
8212 for SDB in response to the @option{-g} option.
8215 @defmac DWARF2_DEBUGGING_INFO
8216 Define this macro if GCC should produce dwarf version 2 format
8217 debugging output in response to the @option{-g} option.
8219 @deftypefn {Target Hook} int TARGET_DWARF_CALLING_CONVENTION (tree @var{function})
8220 Define this to enable the dwarf attribute @code{DW_AT_calling_convention} to
8221 be emitted for each function. Instead of an integer return the enum
8222 value for the @code{DW_CC_} tag.
8225 To support optional call frame debugging information, you must also
8226 define @code{INCOMING_RETURN_ADDR_RTX} and either set
8227 @code{RTX_FRAME_RELATED_P} on the prologue insns if you use RTL for the
8228 prologue, or call @code{dwarf2out_def_cfa} and @code{dwarf2out_reg_save}
8229 as appropriate from @code{TARGET_ASM_FUNCTION_PROLOGUE} if you don't.
8232 @defmac DWARF2_FRAME_INFO
8233 Define this macro to a nonzero value if GCC should always output
8234 Dwarf 2 frame information. If @code{DWARF2_UNWIND_INFO}
8235 (@pxref{Exception Region Output} is nonzero, GCC will output this
8236 information not matter how you define @code{DWARF2_FRAME_INFO}.
8239 @defmac DWARF2_ASM_LINE_DEBUG_INFO
8240 Define this macro to be a nonzero value if the assembler can generate Dwarf 2
8241 line debug info sections. This will result in much more compact line number
8242 tables, and hence is desirable if it works.
8245 @defmac ASM_OUTPUT_DWARF_DELTA (@var{stream}, @var{size}, @var{label1}, @var{label2})
8246 A C statement to issue assembly directives that create a difference
8247 between the two given labels, using an integer of the given size.
8250 @defmac ASM_OUTPUT_DWARF_OFFSET (@var{stream}, @var{size}, @var{label})
8251 A C statement to issue assembly directives that create a
8252 section-relative reference to the given label, using an integer of the
8256 @defmac ASM_OUTPUT_DWARF_PCREL (@var{stream}, @var{size}, @var{label})
8257 A C statement to issue assembly directives that create a self-relative
8258 reference to the given label, using an integer of the given size.
8261 @deftypefn {Target Hook} void TARGET_ASM_OUTPUT_DWARF_DTPREL (FILE *@var{FILE}, int @var{size}, rtx @var{x})
8262 If defined, this target hook is a function which outputs a DTP-relative
8263 reference to the given TLS symbol of the specified size.
8266 @defmac PUT_SDB_@dots{}
8267 Define these macros to override the assembler syntax for the special
8268 SDB assembler directives. See @file{sdbout.c} for a list of these
8269 macros and their arguments. If the standard syntax is used, you need
8270 not define them yourself.
8274 Some assemblers do not support a semicolon as a delimiter, even between
8275 SDB assembler directives. In that case, define this macro to be the
8276 delimiter to use (usually @samp{\n}). It is not necessary to define
8277 a new set of @code{PUT_SDB_@var{op}} macros if this is the only change
8281 @defmac SDB_ALLOW_UNKNOWN_REFERENCES
8282 Define this macro to allow references to unknown structure,
8283 union, or enumeration tags to be emitted. Standard COFF does not
8284 allow handling of unknown references, MIPS ECOFF has support for
8288 @defmac SDB_ALLOW_FORWARD_REFERENCES
8289 Define this macro to allow references to structure, union, or
8290 enumeration tags that have not yet been seen to be handled. Some
8291 assemblers choke if forward tags are used, while some require it.
8294 @defmac SDB_OUTPUT_SOURCE_LINE (@var{stream}, @var{line})
8295 A C statement to output SDB debugging information before code for line
8296 number @var{line} of the current source file to the stdio stream
8297 @var{stream}. The default is to emit an @code{.ln} directive.
8302 @subsection Macros for VMS Debug Format
8304 @c prevent bad page break with this line
8305 Here are macros for VMS debug format.
8307 @defmac VMS_DEBUGGING_INFO
8308 Define this macro if GCC should produce debugging output for VMS
8309 in response to the @option{-g} option. The default behavior for VMS
8310 is to generate minimal debug info for a traceback in the absence of
8311 @option{-g} unless explicitly overridden with @option{-g0}. This
8312 behavior is controlled by @code{OPTIMIZATION_OPTIONS} and
8313 @code{OVERRIDE_OPTIONS}.
8316 @node Floating Point
8317 @section Cross Compilation and Floating Point
8318 @cindex cross compilation and floating point
8319 @cindex floating point and cross compilation
8321 While all modern machines use twos-complement representation for integers,
8322 there are a variety of representations for floating point numbers. This
8323 means that in a cross-compiler the representation of floating point numbers
8324 in the compiled program may be different from that used in the machine
8325 doing the compilation.
8327 Because different representation systems may offer different amounts of
8328 range and precision, all floating point constants must be represented in
8329 the target machine's format. Therefore, the cross compiler cannot
8330 safely use the host machine's floating point arithmetic; it must emulate
8331 the target's arithmetic. To ensure consistency, GCC always uses
8332 emulation to work with floating point values, even when the host and
8333 target floating point formats are identical.
8335 The following macros are provided by @file{real.h} for the compiler to
8336 use. All parts of the compiler which generate or optimize
8337 floating-point calculations must use these macros. They may evaluate
8338 their operands more than once, so operands must not have side effects.
8340 @defmac REAL_VALUE_TYPE
8341 The C data type to be used to hold a floating point value in the target
8342 machine's format. Typically this is a @code{struct} containing an
8343 array of @code{HOST_WIDE_INT}, but all code should treat it as an opaque
8347 @deftypefn Macro int REAL_VALUES_EQUAL (REAL_VALUE_TYPE @var{x}, REAL_VALUE_TYPE @var{y})
8348 Compares for equality the two values, @var{x} and @var{y}. If the target
8349 floating point format supports negative zeroes and/or NaNs,
8350 @samp{REAL_VALUES_EQUAL (-0.0, 0.0)} is true, and
8351 @samp{REAL_VALUES_EQUAL (NaN, NaN)} is false.
8354 @deftypefn Macro int REAL_VALUES_LESS (REAL_VALUE_TYPE @var{x}, REAL_VALUE_TYPE @var{y})
8355 Tests whether @var{x} is less than @var{y}.
8358 @deftypefn Macro HOST_WIDE_INT REAL_VALUE_FIX (REAL_VALUE_TYPE @var{x})
8359 Truncates @var{x} to a signed integer, rounding toward zero.
8362 @deftypefn Macro {unsigned HOST_WIDE_INT} REAL_VALUE_UNSIGNED_FIX (REAL_VALUE_TYPE @var{x})
8363 Truncates @var{x} to an unsigned integer, rounding toward zero. If
8364 @var{x} is negative, returns zero.
8367 @deftypefn Macro REAL_VALUE_TYPE REAL_VALUE_ATOF (const char *@var{string}, enum machine_mode @var{mode})
8368 Converts @var{string} into a floating point number in the target machine's
8369 representation for mode @var{mode}. This routine can handle both
8370 decimal and hexadecimal floating point constants, using the syntax
8371 defined by the C language for both.
8374 @deftypefn Macro int REAL_VALUE_NEGATIVE (REAL_VALUE_TYPE @var{x})
8375 Returns 1 if @var{x} is negative (including negative zero), 0 otherwise.
8378 @deftypefn Macro int REAL_VALUE_ISINF (REAL_VALUE_TYPE @var{x})
8379 Determines whether @var{x} represents infinity (positive or negative).
8382 @deftypefn Macro int REAL_VALUE_ISNAN (REAL_VALUE_TYPE @var{x})
8383 Determines whether @var{x} represents a ``NaN'' (not-a-number).
8386 @deftypefn Macro void REAL_ARITHMETIC (REAL_VALUE_TYPE @var{output}, enum tree_code @var{code}, REAL_VALUE_TYPE @var{x}, REAL_VALUE_TYPE @var{y})
8387 Calculates an arithmetic operation on the two floating point values
8388 @var{x} and @var{y}, storing the result in @var{output} (which must be a
8391 The operation to be performed is specified by @var{code}. Only the
8392 following codes are supported: @code{PLUS_EXPR}, @code{MINUS_EXPR},
8393 @code{MULT_EXPR}, @code{RDIV_EXPR}, @code{MAX_EXPR}, @code{MIN_EXPR}.
8395 If @code{REAL_ARITHMETIC} is asked to evaluate division by zero and the
8396 target's floating point format cannot represent infinity, it will call
8397 @code{abort}. Callers should check for this situation first, using
8398 @code{MODE_HAS_INFINITIES}. @xref{Storage Layout}.
8401 @deftypefn Macro REAL_VALUE_TYPE REAL_VALUE_NEGATE (REAL_VALUE_TYPE @var{x})
8402 Returns the negative of the floating point value @var{x}.
8405 @deftypefn Macro REAL_VALUE_TYPE REAL_VALUE_ABS (REAL_VALUE_TYPE @var{x})
8406 Returns the absolute value of @var{x}.
8409 @deftypefn Macro REAL_VALUE_TYPE REAL_VALUE_TRUNCATE (REAL_VALUE_TYPE @var{mode}, enum machine_mode @var{x})
8410 Truncates the floating point value @var{x} to fit in @var{mode}. The
8411 return value is still a full-size @code{REAL_VALUE_TYPE}, but it has an
8412 appropriate bit pattern to be output asa floating constant whose
8413 precision accords with mode @var{mode}.
8416 @deftypefn Macro void REAL_VALUE_TO_INT (HOST_WIDE_INT @var{low}, HOST_WIDE_INT @var{high}, REAL_VALUE_TYPE @var{x})
8417 Converts a floating point value @var{x} into a double-precision integer
8418 which is then stored into @var{low} and @var{high}. If the value is not
8419 integral, it is truncated.
8422 @deftypefn Macro void REAL_VALUE_FROM_INT (REAL_VALUE_TYPE @var{x}, HOST_WIDE_INT @var{low}, HOST_WIDE_INT @var{high}, enum machine_mode @var{mode})
8423 Converts a double-precision integer found in @var{low} and @var{high},
8424 into a floating point value which is then stored into @var{x}. The
8425 value is truncated to fit in mode @var{mode}.
8428 @node Mode Switching
8429 @section Mode Switching Instructions
8430 @cindex mode switching
8431 The following macros control mode switching optimizations:
8433 @defmac OPTIMIZE_MODE_SWITCHING (@var{entity})
8434 Define this macro if the port needs extra instructions inserted for mode
8435 switching in an optimizing compilation.
8437 For an example, the SH4 can perform both single and double precision
8438 floating point operations, but to perform a single precision operation,
8439 the FPSCR PR bit has to be cleared, while for a double precision
8440 operation, this bit has to be set. Changing the PR bit requires a general
8441 purpose register as a scratch register, hence these FPSCR sets have to
8442 be inserted before reload, i.e.@: you can't put this into instruction emitting
8443 or @code{TARGET_MACHINE_DEPENDENT_REORG}.
8445 You can have multiple entities that are mode-switched, and select at run time
8446 which entities actually need it. @code{OPTIMIZE_MODE_SWITCHING} should
8447 return nonzero for any @var{entity} that needs mode-switching.
8448 If you define this macro, you also have to define
8449 @code{NUM_MODES_FOR_MODE_SWITCHING}, @code{MODE_NEEDED},
8450 @code{MODE_PRIORITY_TO_MODE} and @code{EMIT_MODE_SET}.
8451 @code{MODE_AFTER}, @code{MODE_ENTRY}, and @code{MODE_EXIT}
8455 @defmac NUM_MODES_FOR_MODE_SWITCHING
8456 If you define @code{OPTIMIZE_MODE_SWITCHING}, you have to define this as
8457 initializer for an array of integers. Each initializer element
8458 N refers to an entity that needs mode switching, and specifies the number
8459 of different modes that might need to be set for this entity.
8460 The position of the initializer in the initializer---starting counting at
8461 zero---determines the integer that is used to refer to the mode-switched
8463 In macros that take mode arguments / yield a mode result, modes are
8464 represented as numbers 0 @dots{} N @minus{} 1. N is used to specify that no mode
8465 switch is needed / supplied.
8468 @defmac MODE_NEEDED (@var{entity}, @var{insn})
8469 @var{entity} is an integer specifying a mode-switched entity. If
8470 @code{OPTIMIZE_MODE_SWITCHING} is defined, you must define this macro to
8471 return an integer value not larger than the corresponding element in
8472 @code{NUM_MODES_FOR_MODE_SWITCHING}, to denote the mode that @var{entity} must
8473 be switched into prior to the execution of @var{insn}.
8476 @defmac MODE_AFTER (@var{mode}, @var{insn})
8477 If this macro is defined, it is evaluated for every @var{insn} during
8478 mode switching. It determines the mode that an insn results in (if
8479 different from the incoming mode).
8482 @defmac MODE_ENTRY (@var{entity})
8483 If this macro is defined, it is evaluated for every @var{entity} that needs
8484 mode switching. It should evaluate to an integer, which is a mode that
8485 @var{entity} is assumed to be switched to at function entry. If @code{MODE_ENTRY}
8486 is defined then @code{MODE_EXIT} must be defined.
8489 @defmac MODE_EXIT (@var{entity})
8490 If this macro is defined, it is evaluated for every @var{entity} that needs
8491 mode switching. It should evaluate to an integer, which is a mode that
8492 @var{entity} is assumed to be switched to at function exit. If @code{MODE_EXIT}
8493 is defined then @code{MODE_ENTRY} must be defined.
8496 @defmac MODE_PRIORITY_TO_MODE (@var{entity}, @var{n})
8497 This macro specifies the order in which modes for @var{entity} are processed.
8498 0 is the highest priority, @code{NUM_MODES_FOR_MODE_SWITCHING[@var{entity}] - 1} the
8499 lowest. The value of the macro should be an integer designating a mode
8500 for @var{entity}. For any fixed @var{entity}, @code{mode_priority_to_mode}
8501 (@var{entity}, @var{n}) shall be a bijection in 0 @dots{}
8502 @code{num_modes_for_mode_switching[@var{entity}] - 1}.
8505 @defmac EMIT_MODE_SET (@var{entity}, @var{mode}, @var{hard_regs_live})
8506 Generate one or more insns to set @var{entity} to @var{mode}.
8507 @var{hard_reg_live} is the set of hard registers live at the point where
8508 the insn(s) are to be inserted.
8511 @node Target Attributes
8512 @section Defining target-specific uses of @code{__attribute__}
8513 @cindex target attributes
8514 @cindex machine attributes
8515 @cindex attributes, target-specific
8517 Target-specific attributes may be defined for functions, data and types.
8518 These are described using the following target hooks; they also need to
8519 be documented in @file{extend.texi}.
8521 @deftypevr {Target Hook} {const struct attribute_spec *} TARGET_ATTRIBUTE_TABLE
8522 If defined, this target hook points to an array of @samp{struct
8523 attribute_spec} (defined in @file{tree.h}) specifying the machine
8524 specific attributes for this target and some of the restrictions on the
8525 entities to which these attributes are applied and the arguments they
8529 @deftypefn {Target Hook} int TARGET_COMP_TYPE_ATTRIBUTES (tree @var{type1}, tree @var{type2})
8530 If defined, this target hook is a function which returns zero if the attributes on
8531 @var{type1} and @var{type2} are incompatible, one if they are compatible,
8532 and two if they are nearly compatible (which causes a warning to be
8533 generated). If this is not defined, machine-specific attributes are
8534 supposed always to be compatible.
8537 @deftypefn {Target Hook} void TARGET_SET_DEFAULT_TYPE_ATTRIBUTES (tree @var{type})
8538 If defined, this target hook is a function which assigns default attributes to
8539 newly defined @var{type}.
8542 @deftypefn {Target Hook} tree TARGET_MERGE_TYPE_ATTRIBUTES (tree @var{type1}, tree @var{type2})
8543 Define this target hook if the merging of type attributes needs special
8544 handling. If defined, the result is a list of the combined
8545 @code{TYPE_ATTRIBUTES} of @var{type1} and @var{type2}. It is assumed
8546 that @code{comptypes} has already been called and returned 1. This
8547 function may call @code{merge_attributes} to handle machine-independent
8551 @deftypefn {Target Hook} tree TARGET_MERGE_DECL_ATTRIBUTES (tree @var{olddecl}, tree @var{newdecl})
8552 Define this target hook if the merging of decl attributes needs special
8553 handling. If defined, the result is a list of the combined
8554 @code{DECL_ATTRIBUTES} of @var{olddecl} and @var{newdecl}.
8555 @var{newdecl} is a duplicate declaration of @var{olddecl}. Examples of
8556 when this is needed are when one attribute overrides another, or when an
8557 attribute is nullified by a subsequent definition. This function may
8558 call @code{merge_attributes} to handle machine-independent merging.
8560 @findex TARGET_DLLIMPORT_DECL_ATTRIBUTES
8561 If the only target-specific handling you require is @samp{dllimport}
8562 for Microsoft Windows targets, you should define the macro
8563 @code{TARGET_DLLIMPORT_DECL_ATTRIBUTES} to @code{1}. The compiler
8564 will then define a function called
8565 @code{merge_dllimport_decl_attributes} which can then be defined as
8566 the expansion of @code{TARGET_MERGE_DECL_ATTRIBUTES}. You can also
8567 add @code{handle_dll_attribute} in the attribute table for your port
8568 to perform initial processing of the @samp{dllimport} and
8569 @samp{dllexport} attributes. This is done in @file{i386/cygwin.h} and
8570 @file{i386/i386.c}, for example.
8573 @deftypefn {Target Hook} bool TARGET_VALID_DLLIMPORT_ATTRIBUTE_P (tree @var{decl})
8574 @var{decl} is a variable or function with @code{__attribute__((dllimport))}
8575 specified. Use this hook if the target needs to add extra validation
8576 checks to @code{handle_dll_attribute}.
8579 @defmac TARGET_DECLSPEC
8580 Define this macro to a nonzero value if you want to treat
8581 @code{__declspec(X)} as equivalent to @code{__attribute((X))}. By
8582 default, this behavior is enabled only for targets that define
8583 @code{TARGET_DLLIMPORT_DECL_ATTRIBUTES}. The current implementation
8584 of @code{__declspec} is via a built-in macro, but you should not rely
8585 on this implementation detail.
8588 @deftypefn {Target Hook} void TARGET_INSERT_ATTRIBUTES (tree @var{node}, tree *@var{attr_ptr})
8589 Define this target hook if you want to be able to add attributes to a decl
8590 when it is being created. This is normally useful for back ends which
8591 wish to implement a pragma by using the attributes which correspond to
8592 the pragma's effect. The @var{node} argument is the decl which is being
8593 created. The @var{attr_ptr} argument is a pointer to the attribute list
8594 for this decl. The list itself should not be modified, since it may be
8595 shared with other decls, but attributes may be chained on the head of
8596 the list and @code{*@var{attr_ptr}} modified to point to the new
8597 attributes, or a copy of the list may be made if further changes are
8601 @deftypefn {Target Hook} bool TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P (tree @var{fndecl})
8603 This target hook returns @code{true} if it is ok to inline @var{fndecl}
8604 into the current function, despite its having target-specific
8605 attributes, @code{false} otherwise. By default, if a function has a
8606 target specific attribute attached to it, it will not be inlined.
8609 @node MIPS Coprocessors
8610 @section Defining coprocessor specifics for MIPS targets.
8611 @cindex MIPS coprocessor-definition macros
8613 The MIPS specification allows MIPS implementations to have as many as 4
8614 coprocessors, each with as many as 32 private registers. GCC supports
8615 accessing these registers and transferring values between the registers
8616 and memory using asm-ized variables. For example:
8619 register unsigned int cp0count asm ("c0r1");
8625 (``c0r1'' is the default name of register 1 in coprocessor 0; alternate
8626 names may be added as described below, or the default names may be
8627 overridden entirely in @code{SUBTARGET_CONDITIONAL_REGISTER_USAGE}.)
8629 Coprocessor registers are assumed to be epilogue-used; sets to them will
8630 be preserved even if it does not appear that the register is used again
8631 later in the function.
8633 Another note: according to the MIPS spec, coprocessor 1 (if present) is
8634 the FPU@. One accesses COP1 registers through standard mips
8635 floating-point support; they are not included in this mechanism.
8637 There is one macro used in defining the MIPS coprocessor interface which
8638 you may want to override in subtargets; it is described below.
8640 @defmac ALL_COP_ADDITIONAL_REGISTER_NAMES
8641 A comma-separated list (with leading comma) of pairs describing the
8642 alternate names of coprocessor registers. The format of each entry should be
8644 @{ @var{alternatename}, @var{register_number}@}
8650 @section Parameters for Precompiled Header Validity Checking
8651 @cindex parameters, precompiled headers
8653 @deftypefn {Target Hook} void *TARGET_GET_PCH_VALIDITY (size_t *@var{sz})
8654 This hook returns the data needed by @code{TARGET_PCH_VALID_P} and sets
8655 @samp{*@var{sz}} to the size of the data in bytes.
8658 @deftypefn {Target Hook} const char *TARGET_PCH_VALID_P (const void *@var{data}, size_t @var{sz})
8659 This hook checks whether the options used to create a PCH file are
8660 compatible with the current settings. It returns @code{NULL}
8661 if so and a suitable error message if not. Error messages will
8662 be presented to the user and must be localized using @samp{_(@var{msg})}.
8664 @var{data} is the data that was returned by @code{TARGET_GET_PCH_VALIDITY}
8665 when the PCH file was created and @var{sz} is the size of that data in bytes.
8666 It's safe to assume that the data was created by the same version of the
8667 compiler, so no format checking is needed.
8669 The default definition of @code{default_pch_valid_p} should be
8670 suitable for most targets.
8673 @deftypefn {Target Hook} const char *TARGET_CHECK_PCH_TARGET_FLAGS (int @var{pch_flags})
8674 If this hook is nonnull, the default implementation of
8675 @code{TARGET_PCH_VALID_P} will use it to check for compatible values
8676 of @code{target_flags}. @var{pch_flags} specifies the value that
8677 @code{target_flags} had when the PCH file was created. The return
8678 value is the same as for @code{TARGET_PCH_VALID_P}.
8682 @section C++ ABI parameters
8683 @cindex parameters, c++ abi
8685 @deftypefn {Target Hook} tree TARGET_CXX_GUARD_TYPE (void)
8686 Define this hook to override the integer type used for guard variables.
8687 These are used to implement one-time construction of static objects. The
8688 default is long_long_integer_type_node.
8691 @deftypefn {Target Hook} bool TARGET_CXX_GUARD_MASK_BIT (void)
8692 This hook determines how guard variables are used. It should return
8693 @code{false} (the default) if first byte should be used. A return value of
8694 @code{true} indicates the least significant bit should be used.
8697 @deftypefn {Target Hook} tree TARGET_CXX_GET_COOKIE_SIZE (tree @var{type})
8698 This hook returns the size of the cookie to use when allocating an array
8699 whose elements have the indicated @var{type}. Assumes that it is already
8700 known that a cookie is needed. The default is
8701 @code{max(sizeof (size_t), alignof(type))}, as defined in section 2.7 of the
8702 IA64/Generic C++ ABI@.
8705 @deftypefn {Target Hook} bool TARGET_CXX_COOKIE_HAS_SIZE (void)
8706 This hook should return @code{true} if the element size should be stored in
8707 array cookies. The default is to return @code{false}.
8710 @deftypefn {Target Hook} int TARGET_CXX_IMPORT_EXPORT_CLASS (tree @var{type}, int @var{import_export})
8711 If defined by a backend this hook allows the decision made to export
8712 class @var{type} to be overruled. Upon entry @var{import_export}
8713 will contain 1 if the class is going to be exported, @minus{}1 if it is going
8714 to be imported and 0 otherwise. This function should return the
8715 modified value and perform any other actions necessary to support the
8716 backend's targeted operating system.
8719 @deftypefn {Target Hook} bool TARGET_CXX_CDTOR_RETURNS_THIS (void)
8720 This hook should return @code{true} if constructors and destructors return
8721 the address of the object created/destroyed. The default is to return
8725 @deftypefn {Target Hook} bool TARGET_CXX_KEY_METHOD_MAY_BE_INLINE (void)
8726 This hook returns true if the key method for a class (i.e., the method
8727 which, if defined in the current translation unit, causes the virtual
8728 table to be emitted) may be an inline function. Under the standard
8729 Itanium C++ ABI the key method may be an inline function so long as
8730 the function is not declared inline in the class definition. Under
8731 some variants of the ABI, an inline function can never be the key
8732 method. The default is to return @code{true}.
8735 @deftypefn {Target Hook} void TARGET_CXX_DETERMINE_CLASS_DATA_VISIBILITY (tree @var{decl})
8736 @var{decl} is a virtual table, virtual table table, typeinfo object,
8737 or other similar implicit class data object that will be emitted with
8738 external linkage in this translation unit. No ELF visibility has been
8739 explicitly specified. If the target needs to specify a visibility
8740 other than that of the containing class, use this hook to set
8741 @code{DECL_VISIBILITY} and @code{DECL_VISIBILITY_SPECIFIED}.
8744 @deftypefn {Target Hook} bool TARGET_CXX_CLASS_DATA_ALWAYS_COMDAT (void)
8745 This hook returns true (the default) if virtual tables and other
8746 similar implicit class data objects are always COMDAT if they have
8747 external linkage. If this hook returns false, then class data for
8748 classes whose virtual table will be emitted in only one translation
8749 unit will not be COMDAT.
8752 @deftypefn {Target Hook} bool TARGET_CXX_USE_AEABI_ATEXIT (void)
8753 This hook returns true if @code{__aeabi_atexit} (as defined by the ARM EABI)
8754 should be used to register static destructors when @option{-fuse-cxa-atexit}
8755 is in effect. The default is to return false to use @code{__cxa_atexit}.
8758 @deftypefn {Target Hook} void TARGET_CXX_ADJUST_CLASS_AT_DEFINITION (tree @var{type})
8759 @var{type} is a C++ class (i.e., RECORD_TYPE or UNION_TYPE) that has just been
8760 defined. Use this hook to make adjustments to the class (eg, tweak
8761 visibility or perform any other required target modifications).
8765 @section Miscellaneous Parameters
8766 @cindex parameters, miscellaneous
8768 @c prevent bad page break with this line
8769 Here are several miscellaneous parameters.
8771 @defmac HAS_LONG_COND_BRANCH
8772 Define this boolean macro to indicate whether or not your architecture
8773 has conditional branches that can span all of memory. It is used in
8774 conjunction with an optimization that partitions hot and cold basic
8775 blocks into separate sections of the executable. If this macro is
8776 set to false, gcc will convert any conditional branches that attempt
8777 to cross between sections into unconditional branches or indirect jumps.
8780 @defmac HAS_LONG_UNCOND_BRANCH
8781 Define this boolean macro to indicate whether or not your architecture
8782 has unconditional branches that can span all of memory. It is used in
8783 conjunction with an optimization that partitions hot and cold basic
8784 blocks into separate sections of the executable. If this macro is
8785 set to false, gcc will convert any unconditional branches that attempt
8786 to cross between sections into indirect jumps.
8789 @defmac CASE_VECTOR_MODE
8790 An alias for a machine mode name. This is the machine mode that
8791 elements of a jump-table should have.
8794 @defmac CASE_VECTOR_SHORTEN_MODE (@var{min_offset}, @var{max_offset}, @var{body})
8795 Optional: return the preferred mode for an @code{addr_diff_vec}
8796 when the minimum and maximum offset are known. If you define this,
8797 it enables extra code in branch shortening to deal with @code{addr_diff_vec}.
8798 To make this work, you also have to define @code{INSN_ALIGN} and
8799 make the alignment for @code{addr_diff_vec} explicit.
8800 The @var{body} argument is provided so that the offset_unsigned and scale
8801 flags can be updated.
8804 @defmac CASE_VECTOR_PC_RELATIVE
8805 Define this macro to be a C expression to indicate when jump-tables
8806 should contain relative addresses. You need not define this macro if
8807 jump-tables never contain relative addresses, or jump-tables should
8808 contain relative addresses only when @option{-fPIC} or @option{-fPIC}
8812 @defmac CASE_VALUES_THRESHOLD
8813 Define this to be the smallest number of different values for which it
8814 is best to use a jump-table instead of a tree of conditional branches.
8815 The default is four for machines with a @code{casesi} instruction and
8816 five otherwise. This is best for most machines.
8819 @defmac CASE_USE_BIT_TESTS
8820 Define this macro to be a C expression to indicate whether C switch
8821 statements may be implemented by a sequence of bit tests. This is
8822 advantageous on processors that can efficiently implement left shift
8823 of 1 by the number of bits held in a register, but inappropriate on
8824 targets that would require a loop. By default, this macro returns
8825 @code{true} if the target defines an @code{ashlsi3} pattern, and
8826 @code{false} otherwise.
8829 @defmac WORD_REGISTER_OPERATIONS
8830 Define this macro if operations between registers with integral mode
8831 smaller than a word are always performed on the entire register.
8832 Most RISC machines have this property and most CISC machines do not.
8835 @defmac LOAD_EXTEND_OP (@var{mem_mode})
8836 Define this macro to be a C expression indicating when insns that read
8837 memory in @var{mem_mode}, an integral mode narrower than a word, set the
8838 bits outside of @var{mem_mode} to be either the sign-extension or the
8839 zero-extension of the data read. Return @code{SIGN_EXTEND} for values
8840 of @var{mem_mode} for which the
8841 insn sign-extends, @code{ZERO_EXTEND} for which it zero-extends, and
8842 @code{UNKNOWN} for other modes.
8844 This macro is not called with @var{mem_mode} non-integral or with a width
8845 greater than or equal to @code{BITS_PER_WORD}, so you may return any
8846 value in this case. Do not define this macro if it would always return
8847 @code{UNKNOWN}. On machines where this macro is defined, you will normally
8848 define it as the constant @code{SIGN_EXTEND} or @code{ZERO_EXTEND}.
8850 You may return a non-@code{UNKNOWN} value even if for some hard registers
8851 the sign extension is not performed, if for the @code{REGNO_REG_CLASS}
8852 of these hard registers @code{CANNOT_CHANGE_MODE_CLASS} returns nonzero
8853 when the @var{from} mode is @var{mem_mode} and the @var{to} mode is any
8854 integral mode larger than this but not larger than @code{word_mode}.
8856 You must return @code{UNKNOWN} if for some hard registers that allow this
8857 mode, @code{CANNOT_CHANGE_MODE_CLASS} says that they cannot change to
8858 @code{word_mode}, but that they can change to another integral mode that
8859 is larger then @var{mem_mode} but still smaller than @code{word_mode}.
8862 @defmac SHORT_IMMEDIATES_SIGN_EXTEND
8863 Define this macro if loading short immediate values into registers sign
8867 @defmac FIXUNS_TRUNC_LIKE_FIX_TRUNC
8868 Define this macro if the same instructions that convert a floating
8869 point number to a signed fixed point number also convert validly to an
8874 The maximum number of bytes that a single instruction can move quickly
8875 between memory and registers or between two memory locations.
8878 @defmac MAX_MOVE_MAX
8879 The maximum number of bytes that a single instruction can move quickly
8880 between memory and registers or between two memory locations. If this
8881 is undefined, the default is @code{MOVE_MAX}. Otherwise, it is the
8882 constant value that is the largest value that @code{MOVE_MAX} can have
8886 @defmac SHIFT_COUNT_TRUNCATED
8887 A C expression that is nonzero if on this machine the number of bits
8888 actually used for the count of a shift operation is equal to the number
8889 of bits needed to represent the size of the object being shifted. When
8890 this macro is nonzero, the compiler will assume that it is safe to omit
8891 a sign-extend, zero-extend, and certain bitwise `and' instructions that
8892 truncates the count of a shift operation. On machines that have
8893 instructions that act on bit-fields at variable positions, which may
8894 include `bit test' instructions, a nonzero @code{SHIFT_COUNT_TRUNCATED}
8895 also enables deletion of truncations of the values that serve as
8896 arguments to bit-field instructions.
8898 If both types of instructions truncate the count (for shifts) and
8899 position (for bit-field operations), or if no variable-position bit-field
8900 instructions exist, you should define this macro.
8902 However, on some machines, such as the 80386 and the 680x0, truncation
8903 only applies to shift operations and not the (real or pretended)
8904 bit-field operations. Define @code{SHIFT_COUNT_TRUNCATED} to be zero on
8905 such machines. Instead, add patterns to the @file{md} file that include
8906 the implied truncation of the shift instructions.
8908 You need not define this macro if it would always have the value of zero.
8911 @anchor{TARGET_SHIFT_TRUNCATION_MASK}
8912 @deftypefn {Target Hook} int TARGET_SHIFT_TRUNCATION_MASK (enum machine_mode @var{mode})
8913 This function describes how the standard shift patterns for @var{mode}
8914 deal with shifts by negative amounts or by more than the width of the mode.
8915 @xref{shift patterns}.
8917 On many machines, the shift patterns will apply a mask @var{m} to the
8918 shift count, meaning that a fixed-width shift of @var{x} by @var{y} is
8919 equivalent to an arbitrary-width shift of @var{x} by @var{y & m}. If
8920 this is true for mode @var{mode}, the function should return @var{m},
8921 otherwise it should return 0. A return value of 0 indicates that no
8922 particular behavior is guaranteed.
8924 Note that, unlike @code{SHIFT_COUNT_TRUNCATED}, this function does
8925 @emph{not} apply to general shift rtxes; it applies only to instructions
8926 that are generated by the named shift patterns.
8928 The default implementation of this function returns
8929 @code{GET_MODE_BITSIZE (@var{mode}) - 1} if @code{SHIFT_COUNT_TRUNCATED}
8930 and 0 otherwise. This definition is always safe, but if
8931 @code{SHIFT_COUNT_TRUNCATED} is false, and some shift patterns
8932 nevertheless truncate the shift count, you may get better code
8936 @defmac TRULY_NOOP_TRUNCATION (@var{outprec}, @var{inprec})
8937 A C expression which is nonzero if on this machine it is safe to
8938 ``convert'' an integer of @var{inprec} bits to one of @var{outprec}
8939 bits (where @var{outprec} is smaller than @var{inprec}) by merely
8940 operating on it as if it had only @var{outprec} bits.
8942 On many machines, this expression can be 1.
8944 @c rearranged this, removed the phrase "it is reported that". this was
8945 @c to fix an overfull hbox. --mew 10feb93
8946 When @code{TRULY_NOOP_TRUNCATION} returns 1 for a pair of sizes for
8947 modes for which @code{MODES_TIEABLE_P} is 0, suboptimal code can result.
8948 If this is the case, making @code{TRULY_NOOP_TRUNCATION} return 0 in
8949 such cases may improve things.
8952 @defmac STORE_FLAG_VALUE
8953 A C expression describing the value returned by a comparison operator
8954 with an integral mode and stored by a store-flag instruction
8955 (@samp{s@var{cond}}) when the condition is true. This description must
8956 apply to @emph{all} the @samp{s@var{cond}} patterns and all the
8957 comparison operators whose results have a @code{MODE_INT} mode.
8959 A value of 1 or @minus{}1 means that the instruction implementing the
8960 comparison operator returns exactly 1 or @minus{}1 when the comparison is true
8961 and 0 when the comparison is false. Otherwise, the value indicates
8962 which bits of the result are guaranteed to be 1 when the comparison is
8963 true. This value is interpreted in the mode of the comparison
8964 operation, which is given by the mode of the first operand in the
8965 @samp{s@var{cond}} pattern. Either the low bit or the sign bit of
8966 @code{STORE_FLAG_VALUE} be on. Presently, only those bits are used by
8969 If @code{STORE_FLAG_VALUE} is neither 1 or @minus{}1, the compiler will
8970 generate code that depends only on the specified bits. It can also
8971 replace comparison operators with equivalent operations if they cause
8972 the required bits to be set, even if the remaining bits are undefined.
8973 For example, on a machine whose comparison operators return an
8974 @code{SImode} value and where @code{STORE_FLAG_VALUE} is defined as
8975 @samp{0x80000000}, saying that just the sign bit is relevant, the
8979 (ne:SI (and:SI @var{x} (const_int @var{power-of-2})) (const_int 0))
8986 (ashift:SI @var{x} (const_int @var{n}))
8990 where @var{n} is the appropriate shift count to move the bit being
8991 tested into the sign bit.
8993 There is no way to describe a machine that always sets the low-order bit
8994 for a true value, but does not guarantee the value of any other bits,
8995 but we do not know of any machine that has such an instruction. If you
8996 are trying to port GCC to such a machine, include an instruction to
8997 perform a logical-and of the result with 1 in the pattern for the
8998 comparison operators and let us know at @email{gcc@@gcc.gnu.org}.
9000 Often, a machine will have multiple instructions that obtain a value
9001 from a comparison (or the condition codes). Here are rules to guide the
9002 choice of value for @code{STORE_FLAG_VALUE}, and hence the instructions
9007 Use the shortest sequence that yields a valid definition for
9008 @code{STORE_FLAG_VALUE}. It is more efficient for the compiler to
9009 ``normalize'' the value (convert it to, e.g., 1 or 0) than for the
9010 comparison operators to do so because there may be opportunities to
9011 combine the normalization with other operations.
9014 For equal-length sequences, use a value of 1 or @minus{}1, with @minus{}1 being
9015 slightly preferred on machines with expensive jumps and 1 preferred on
9019 As a second choice, choose a value of @samp{0x80000001} if instructions
9020 exist that set both the sign and low-order bits but do not define the
9024 Otherwise, use a value of @samp{0x80000000}.
9027 Many machines can produce both the value chosen for
9028 @code{STORE_FLAG_VALUE} and its negation in the same number of
9029 instructions. On those machines, you should also define a pattern for
9030 those cases, e.g., one matching
9033 (set @var{A} (neg:@var{m} (ne:@var{m} @var{B} @var{C})))
9036 Some machines can also perform @code{and} or @code{plus} operations on
9037 condition code values with less instructions than the corresponding
9038 @samp{s@var{cond}} insn followed by @code{and} or @code{plus}. On those
9039 machines, define the appropriate patterns. Use the names @code{incscc}
9040 and @code{decscc}, respectively, for the patterns which perform
9041 @code{plus} or @code{minus} operations on condition code values. See
9042 @file{rs6000.md} for some examples. The GNU Superoptizer can be used to
9043 find such instruction sequences on other machines.
9045 If this macro is not defined, the default value, 1, is used. You need
9046 not define @code{STORE_FLAG_VALUE} if the machine has no store-flag
9047 instructions, or if the value generated by these instructions is 1.
9050 @defmac FLOAT_STORE_FLAG_VALUE (@var{mode})
9051 A C expression that gives a nonzero @code{REAL_VALUE_TYPE} value that is
9052 returned when comparison operators with floating-point results are true.
9053 Define this macro on machines that have comparison operations that return
9054 floating-point values. If there are no such operations, do not define
9058 @defmac VECTOR_STORE_FLAG_VALUE (@var{mode})
9059 A C expression that gives a rtx representing the nonzero true element
9060 for vector comparisons. The returned rtx should be valid for the inner
9061 mode of @var{mode} which is guaranteed to be a vector mode. Define
9062 this macro on machines that have vector comparison operations that
9063 return a vector result. If there are no such operations, do not define
9064 this macro. Typically, this macro is defined as @code{const1_rtx} or
9065 @code{constm1_rtx}. This macro may return @code{NULL_RTX} to prevent
9066 the compiler optimizing such vector comparison operations for the
9070 @defmac CLZ_DEFINED_VALUE_AT_ZERO (@var{mode}, @var{value})
9071 @defmacx CTZ_DEFINED_VALUE_AT_ZERO (@var{mode}, @var{value})
9072 A C expression that evaluates to true if the architecture defines a value
9073 for @code{clz} or @code{ctz} with a zero operand. If so, @var{value}
9074 should be set to this value. If this macro is not defined, the value of
9075 @code{clz} or @code{ctz} is assumed to be undefined.
9077 This macro must be defined if the target's expansion for @code{ffs}
9078 relies on a particular value to get correct results. Otherwise it
9079 is not necessary, though it may be used to optimize some corner cases.
9081 Note that regardless of this macro the ``definedness'' of @code{clz}
9082 and @code{ctz} at zero do @emph{not} extend to the builtin functions
9083 visible to the user. Thus one may be free to adjust the value at will
9084 to match the target expansion of these operations without fear of
9089 An alias for the machine mode for pointers. On most machines, define
9090 this to be the integer mode corresponding to the width of a hardware
9091 pointer; @code{SImode} on 32-bit machine or @code{DImode} on 64-bit machines.
9092 On some machines you must define this to be one of the partial integer
9093 modes, such as @code{PSImode}.
9095 The width of @code{Pmode} must be at least as large as the value of
9096 @code{POINTER_SIZE}. If it is not equal, you must define the macro
9097 @code{POINTERS_EXTEND_UNSIGNED} to specify how pointers are extended
9101 @defmac FUNCTION_MODE
9102 An alias for the machine mode used for memory references to functions
9103 being called, in @code{call} RTL expressions. On most machines this
9104 should be @code{QImode}.
9107 @defmac STDC_0_IN_SYSTEM_HEADERS
9108 In normal operation, the preprocessor expands @code{__STDC__} to the
9109 constant 1, to signify that GCC conforms to ISO Standard C@. On some
9110 hosts, like Solaris, the system compiler uses a different convention,
9111 where @code{__STDC__} is normally 0, but is 1 if the user specifies
9112 strict conformance to the C Standard.
9114 Defining @code{STDC_0_IN_SYSTEM_HEADERS} makes GNU CPP follows the host
9115 convention when processing system header files, but when processing user
9116 files @code{__STDC__} will always expand to 1.
9119 @defmac NO_IMPLICIT_EXTERN_C
9120 Define this macro if the system header files support C++ as well as C@.
9121 This macro inhibits the usual method of using system header files in
9122 C++, which is to pretend that the file's contents are enclosed in
9123 @samp{extern "C" @{@dots{}@}}.
9128 @defmac REGISTER_TARGET_PRAGMAS ()
9129 Define this macro if you want to implement any target-specific pragmas.
9130 If defined, it is a C expression which makes a series of calls to
9131 @code{c_register_pragma} or @code{c_register_pragma_with_expansion}
9132 for each pragma. The macro may also do any
9133 setup required for the pragmas.
9135 The primary reason to define this macro is to provide compatibility with
9136 other compilers for the same target. In general, we discourage
9137 definition of target-specific pragmas for GCC@.
9139 If the pragma can be implemented by attributes then you should consider
9140 defining the target hook @samp{TARGET_INSERT_ATTRIBUTES} as well.
9142 Preprocessor macros that appear on pragma lines are not expanded. All
9143 @samp{#pragma} directives that do not match any registered pragma are
9144 silently ignored, unless the user specifies @option{-Wunknown-pragmas}.
9147 @deftypefun void c_register_pragma (const char *@var{space}, const char *@var{name}, void (*@var{callback}) (struct cpp_reader *))
9148 @deftypefunx void c_register_pragma_with_expansion (const char *@var{space}, const char *@var{name}, void (*@var{callback}) (struct cpp_reader *))
9150 Each call to @code{c_register_pragma} or
9151 @code{c_register_pragma_with_expansion} establishes one pragma. The
9152 @var{callback} routine will be called when the preprocessor encounters a
9156 #pragma [@var{space}] @var{name} @dots{}
9159 @var{space} is the case-sensitive namespace of the pragma, or
9160 @code{NULL} to put the pragma in the global namespace. The callback
9161 routine receives @var{pfile} as its first argument, which can be passed
9162 on to cpplib's functions if necessary. You can lex tokens after the
9163 @var{name} by calling @code{pragma_lex}. Tokens that are not read by the
9164 callback will be silently ignored. The end of the line is indicated by
9165 a token of type @code{CPP_EOF}. Macro expansion occurs on the
9166 arguments of pragmas registered with
9167 @code{c_register_pragma_with_expansion} but not on the arguments of
9168 pragmas registered with @code{c_register_pragma}.
9170 For an example use of this routine, see @file{c4x.h} and the callback
9171 routines defined in @file{c4x-c.c}.
9173 Note that the use of @code{pragma_lex} is specific to the C and C++
9174 compilers. It will not work in the Java or Fortran compilers, or any
9175 other language compilers for that matter. Thus if @code{pragma_lex} is going
9176 to be called from target-specific code, it must only be done so when
9177 building the C and C++ compilers. This can be done by defining the
9178 variables @code{c_target_objs} and @code{cxx_target_objs} in the
9179 target entry in the @file{config.gcc} file. These variables should name
9180 the target-specific, language-specific object file which contains the
9181 code that uses @code{pragma_lex}. Note it will also be necessary to add a
9182 rule to the makefile fragment pointed to by @code{tmake_file} that shows
9183 how to build this object file.
9188 @defmac HANDLE_SYSV_PRAGMA
9189 Define this macro (to a value of 1) if you want the System V style
9190 pragmas @samp{#pragma pack(<n>)} and @samp{#pragma weak <name>
9191 [=<value>]} to be supported by gcc.
9193 The pack pragma specifies the maximum alignment (in bytes) of fields
9194 within a structure, in much the same way as the @samp{__aligned__} and
9195 @samp{__packed__} @code{__attribute__}s do. A pack value of zero resets
9196 the behavior to the default.
9198 A subtlety for Microsoft Visual C/C++ style bit-field packing
9199 (e.g.@: -mms-bitfields) for targets that support it:
9200 When a bit-field is inserted into a packed record, the whole size
9201 of the underlying type is used by one or more same-size adjacent
9202 bit-fields (that is, if its long:3, 32 bits is used in the record,
9203 and any additional adjacent long bit-fields are packed into the same
9204 chunk of 32 bits. However, if the size changes, a new field of that
9207 If both MS bit-fields and @samp{__attribute__((packed))} are used,
9208 the latter will take precedence. If @samp{__attribute__((packed))} is
9209 used on a single field when MS bit-fields are in use, it will take
9210 precedence for that field, but the alignment of the rest of the structure
9211 may affect its placement.
9213 The weak pragma only works if @code{SUPPORTS_WEAK} and
9214 @code{ASM_WEAKEN_LABEL} are defined. If enabled it allows the creation
9215 of specifically named weak labels, optionally with a value.
9220 @defmac HANDLE_PRAGMA_PACK_PUSH_POP
9221 Define this macro (to a value of 1) if you want to support the Win32
9222 style pragmas @samp{#pragma pack(push[,@var{n}])} and @samp{#pragma
9223 pack(pop)}. The @samp{pack(push,[@var{n}])} pragma specifies the maximum
9224 alignment (in bytes) of fields within a structure, in much the same way as
9225 the @samp{__aligned__} and @samp{__packed__} @code{__attribute__}s do. A
9226 pack value of zero resets the behavior to the default. Successive
9227 invocations of this pragma cause the previous values to be stacked, so
9228 that invocations of @samp{#pragma pack(pop)} will return to the previous
9232 @defmac HANDLE_PRAGMA_PACK_WITH_EXPANSION
9233 Define this macro, as well as
9234 @code{HANDLE_SYSV_PRAGMA}, if macros should be expanded in the
9235 arguments of @samp{#pragma pack}.
9238 @defmac TARGET_DEFAULT_PACK_STRUCT
9239 If your target requires a structure packing default other than 0 (meaning
9240 the machine default), define this macro to the necessary value (in bytes).
9241 This must be a value that would also valid to be used with
9242 @samp{#pragma pack()} (that is, a small power of two).
9245 @defmac DOLLARS_IN_IDENTIFIERS
9246 Define this macro to control use of the character @samp{$} in
9247 identifier names for the C family of languages. 0 means @samp{$} is
9248 not allowed by default; 1 means it is allowed. 1 is the default;
9249 there is no need to define this macro in that case.
9252 @defmac NO_DOLLAR_IN_LABEL
9253 Define this macro if the assembler does not accept the character
9254 @samp{$} in label names. By default constructors and destructors in
9255 G++ have @samp{$} in the identifiers. If this macro is defined,
9256 @samp{.} is used instead.
9259 @defmac NO_DOT_IN_LABEL
9260 Define this macro if the assembler does not accept the character
9261 @samp{.} in label names. By default constructors and destructors in G++
9262 have names that use @samp{.}. If this macro is defined, these names
9263 are rewritten to avoid @samp{.}.
9266 @defmac INSN_SETS_ARE_DELAYED (@var{insn})
9267 Define this macro as a C expression that is nonzero if it is safe for the
9268 delay slot scheduler to place instructions in the delay slot of @var{insn},
9269 even if they appear to use a resource set or clobbered in @var{insn}.
9270 @var{insn} is always a @code{jump_insn} or an @code{insn}; GCC knows that
9271 every @code{call_insn} has this behavior. On machines where some @code{insn}
9272 or @code{jump_insn} is really a function call and hence has this behavior,
9273 you should define this macro.
9275 You need not define this macro if it would always return zero.
9278 @defmac INSN_REFERENCES_ARE_DELAYED (@var{insn})
9279 Define this macro as a C expression that is nonzero if it is safe for the
9280 delay slot scheduler to place instructions in the delay slot of @var{insn},
9281 even if they appear to set or clobber a resource referenced in @var{insn}.
9282 @var{insn} is always a @code{jump_insn} or an @code{insn}. On machines where
9283 some @code{insn} or @code{jump_insn} is really a function call and its operands
9284 are registers whose use is actually in the subroutine it calls, you should
9285 define this macro. Doing so allows the delay slot scheduler to move
9286 instructions which copy arguments into the argument registers into the delay
9289 You need not define this macro if it would always return zero.
9292 @defmac MULTIPLE_SYMBOL_SPACES
9293 Define this macro as a C expression that is nonzero if, in some cases,
9294 global symbols from one translation unit may not be bound to undefined
9295 symbols in another translation unit without user intervention. For
9296 instance, under Microsoft Windows symbols must be explicitly imported
9297 from shared libraries (DLLs).
9299 You need not define this macro if it would always evaluate to zero.
9302 @deftypefn {Target Hook} tree TARGET_MD_ASM_CLOBBERS (tree @var{outputs}, tree @var{inputs}, tree @var{clobbers})
9303 This target hook should add to @var{clobbers} @code{STRING_CST} trees for
9304 any hard regs the port wishes to automatically clobber for an asm.
9305 It should return the result of the last @code{tree_cons} used to add a
9306 clobber. The @var{outputs}, @var{inputs} and @var{clobber} lists are the
9307 corresponding parameters to the asm and may be inspected to avoid
9308 clobbering a register that is an input or output of the asm. You can use
9309 @code{decl_overlaps_hard_reg_set_p}, declared in @file{tree.h}, to test
9310 for overlap with regards to asm-declared registers.
9313 @defmac MATH_LIBRARY
9314 Define this macro as a C string constant for the linker argument to link
9315 in the system math library, or @samp{""} if the target does not have a
9316 separate math library.
9318 You need only define this macro if the default of @samp{"-lm"} is wrong.
9321 @defmac LIBRARY_PATH_ENV
9322 Define this macro as a C string constant for the environment variable that
9323 specifies where the linker should look for libraries.
9325 You need only define this macro if the default of @samp{"LIBRARY_PATH"}
9329 @defmac TARGET_POSIX_IO
9330 Define this macro if the target supports the following POSIX@ file
9331 functions, access, mkdir and file locking with fcntl / F_SETLKW@.
9332 Defining @code{TARGET_POSIX_IO} will enable the test coverage code
9333 to use file locking when exiting a program, which avoids race conditions
9334 if the program has forked. It will also create directories at run-time
9335 for cross-profiling.
9338 @defmac MAX_CONDITIONAL_EXECUTE
9340 A C expression for the maximum number of instructions to execute via
9341 conditional execution instructions instead of a branch. A value of
9342 @code{BRANCH_COST}+1 is the default if the machine does not use cc0, and
9343 1 if it does use cc0.
9346 @defmac IFCVT_MODIFY_TESTS (@var{ce_info}, @var{true_expr}, @var{false_expr})
9347 Used if the target needs to perform machine-dependent modifications on the
9348 conditionals used for turning basic blocks into conditionally executed code.
9349 @var{ce_info} points to a data structure, @code{struct ce_if_block}, which
9350 contains information about the currently processed blocks. @var{true_expr}
9351 and @var{false_expr} are the tests that are used for converting the
9352 then-block and the else-block, respectively. Set either @var{true_expr} or
9353 @var{false_expr} to a null pointer if the tests cannot be converted.
9356 @defmac IFCVT_MODIFY_MULTIPLE_TESTS (@var{ce_info}, @var{bb}, @var{true_expr}, @var{false_expr})
9357 Like @code{IFCVT_MODIFY_TESTS}, but used when converting more complicated
9358 if-statements into conditions combined by @code{and} and @code{or} operations.
9359 @var{bb} contains the basic block that contains the test that is currently
9360 being processed and about to be turned into a condition.
9363 @defmac IFCVT_MODIFY_INSN (@var{ce_info}, @var{pattern}, @var{insn})
9364 A C expression to modify the @var{PATTERN} of an @var{INSN} that is to
9365 be converted to conditional execution format. @var{ce_info} points to
9366 a data structure, @code{struct ce_if_block}, which contains information
9367 about the currently processed blocks.
9370 @defmac IFCVT_MODIFY_FINAL (@var{ce_info})
9371 A C expression to perform any final machine dependent modifications in
9372 converting code to conditional execution. The involved basic blocks
9373 can be found in the @code{struct ce_if_block} structure that is pointed
9374 to by @var{ce_info}.
9377 @defmac IFCVT_MODIFY_CANCEL (@var{ce_info})
9378 A C expression to cancel any machine dependent modifications in
9379 converting code to conditional execution. The involved basic blocks
9380 can be found in the @code{struct ce_if_block} structure that is pointed
9381 to by @var{ce_info}.
9384 @defmac IFCVT_INIT_EXTRA_FIELDS (@var{ce_info})
9385 A C expression to initialize any extra fields in a @code{struct ce_if_block}
9386 structure, which are defined by the @code{IFCVT_EXTRA_FIELDS} macro.
9389 @defmac IFCVT_EXTRA_FIELDS
9390 If defined, it should expand to a set of field declarations that will be
9391 added to the @code{struct ce_if_block} structure. These should be initialized
9392 by the @code{IFCVT_INIT_EXTRA_FIELDS} macro.
9395 @deftypefn {Target Hook} void TARGET_MACHINE_DEPENDENT_REORG ()
9396 If non-null, this hook performs a target-specific pass over the
9397 instruction stream. The compiler will run it at all optimization levels,
9398 just before the point at which it normally does delayed-branch scheduling.
9400 The exact purpose of the hook varies from target to target. Some use
9401 it to do transformations that are necessary for correctness, such as
9402 laying out in-function constant pools or avoiding hardware hazards.
9403 Others use it as an opportunity to do some machine-dependent optimizations.
9405 You need not implement the hook if it has nothing to do. The default
9409 @deftypefn {Target Hook} void TARGET_INIT_BUILTINS ()
9410 Define this hook if you have any machine-specific built-in functions
9411 that need to be defined. It should be a function that performs the
9414 Machine specific built-in functions can be useful to expand special machine
9415 instructions that would otherwise not normally be generated because
9416 they have no equivalent in the source language (for example, SIMD vector
9417 instructions or prefetch instructions).
9419 To create a built-in function, call the function
9420 @code{lang_hooks.builtin_function}
9421 which is defined by the language front end. You can use any type nodes set
9422 up by @code{build_common_tree_nodes} and @code{build_common_tree_nodes_2};
9423 only language front ends that use those two functions will call
9424 @samp{TARGET_INIT_BUILTINS}.
9427 @deftypefn {Target Hook} rtx TARGET_EXPAND_BUILTIN (tree @var{exp}, rtx @var{target}, rtx @var{subtarget}, enum machine_mode @var{mode}, int @var{ignore})
9429 Expand a call to a machine specific built-in function that was set up by
9430 @samp{TARGET_INIT_BUILTINS}. @var{exp} is the expression for the
9431 function call; the result should go to @var{target} if that is
9432 convenient, and have mode @var{mode} if that is convenient.
9433 @var{subtarget} may be used as the target for computing one of
9434 @var{exp}'s operands. @var{ignore} is nonzero if the value is to be
9435 ignored. This function should return the result of the call to the
9439 @deftypefn {Target Hook} tree TARGET_RESOLVE_OVERLOADED_BUILTIN (tree @var{fndecl}, tree @var{arglist})
9441 Select a replacement for a machine specific built-in function that
9442 was set up by @samp{TARGET_INIT_BUILTINS}. This is done
9443 @emph{before} regular type checking, and so allows the target to
9444 implement a crude form of function overloading. @var{fndecl} is the
9445 declaration of the built-in function. @var{arglist} is the list of
9446 arguments passed to the built-in function. The result is a
9447 complete expression that implements the operation, usually
9448 another @code{CALL_EXPR}.
9451 @deftypefn {Target Hook} tree TARGET_FOLD_BUILTIN (tree @var{fndecl}, tree @var{arglist}, bool @var{ignore})
9453 Fold a call to a machine specific built-in function that was set up by
9454 @samp{TARGET_INIT_BUILTINS}. @var{fndecl} is the declaration of the
9455 built-in function. @var{arglist} is the list of arguments passed to
9456 the built-in function. The result is another tree containing a
9457 simplified expression for the call's result. If @var{ignore} is true
9458 the value will be ignored.
9461 @deftypefn {Target Hook} const char * TARGET_INVALID_WITHIN_DOLOOP (rtx @var{insn})
9463 Take an instruction in @var{insn} and return NULL if it is valid within a
9464 low-overhead loop, otherwise return a string why doloop could not be applied.
9466 Many targets use special registers for low-overhead looping. For any
9467 instruction that clobbers these this function should return a string indicating
9468 the reason why the doloop could not be applied.
9469 By default, the RTL loop optimizer does not use a present doloop pattern for
9470 loops containing function calls or branch on table instructions.
9473 @defmac MD_CAN_REDIRECT_BRANCH (@var{branch1}, @var{branch2})
9475 Take a branch insn in @var{branch1} and another in @var{branch2}.
9476 Return true if redirecting @var{branch1} to the destination of
9477 @var{branch2} is possible.
9479 On some targets, branches may have a limited range. Optimizing the
9480 filling of delay slots can result in branches being redirected, and this
9481 may in turn cause a branch offset to overflow.
9484 @deftypefn {Target Hook} bool TARGET_COMMUTATIVE_P (rtx @var{x}, @var{outer_code})
9485 This target hook returns @code{true} if @var{x} is considered to be commutative.
9486 Usually, this is just COMMUTATIVE_P (@var{x}), but the HP PA doesn't consider
9487 PLUS to be commutative inside a MEM. @var{outer_code} is the rtx code
9488 of the enclosing rtl, if known, otherwise it is UNKNOWN.
9491 @deftypefn {Target Hook} rtx TARGET_ALLOCATE_INITIAL_VALUE (rtx @var{hard_reg})
9493 When the initial value of a hard register has been copied in a pseudo
9494 register, it is often not necessary to actually allocate another register
9495 to this pseudo register, because the original hard register or a stack slot
9496 it has been saved into can be used. @code{TARGET_ALLOCATE_INITIAL_VALUE}
9497 is called at the start of register allocation once for each hard register
9498 that had its initial value copied by using
9499 @code{get_func_hard_reg_initial_val} or @code{get_hard_reg_initial_val}.
9500 Possible values are @code{NULL_RTX}, if you don't want
9501 to do any special allocation, a @code{REG} rtx---that would typically be
9502 the hard register itself, if it is known not to be clobbered---or a
9504 If you are returning a @code{MEM}, this is only a hint for the allocator;
9505 it might decide to use another register anyways.
9506 You may use @code{current_function_leaf_function} in the hook, functions
9507 that use @code{REG_N_SETS}, to determine if the hard
9508 register in question will not be clobbered.
9509 The default value of this hook is @code{NULL}, which disables any special
9513 @defmac TARGET_OBJECT_SUFFIX
9514 Define this macro to be a C string representing the suffix for object
9515 files on your target machine. If you do not define this macro, GCC will
9516 use @samp{.o} as the suffix for object files.
9519 @defmac TARGET_EXECUTABLE_SUFFIX
9520 Define this macro to be a C string representing the suffix to be
9521 automatically added to executable files on your target machine. If you
9522 do not define this macro, GCC will use the null string as the suffix for
9526 @defmac COLLECT_EXPORT_LIST
9527 If defined, @code{collect2} will scan the individual object files
9528 specified on its command line and create an export list for the linker.
9529 Define this macro for systems like AIX, where the linker discards
9530 object files that are not referenced from @code{main} and uses export
9534 @defmac MODIFY_JNI_METHOD_CALL (@var{mdecl})
9535 Define this macro to a C expression representing a variant of the
9536 method call @var{mdecl}, if Java Native Interface (JNI) methods
9537 must be invoked differently from other methods on your target.
9538 For example, on 32-bit Microsoft Windows, JNI methods must be invoked using
9539 the @code{stdcall} calling convention and this macro is then
9540 defined as this expression:
9543 build_type_attribute_variant (@var{mdecl},
9545 (get_identifier ("stdcall"),
9550 @deftypefn {Target Hook} bool TARGET_CANNOT_MODIFY_JUMPS_P (void)
9551 This target hook returns @code{true} past the point in which new jump
9552 instructions could be created. On machines that require a register for
9553 every jump such as the SHmedia ISA of SH5, this point would typically be
9554 reload, so this target hook should be defined to a function such as:
9558 cannot_modify_jumps_past_reload_p ()
9560 return (reload_completed || reload_in_progress);
9565 @deftypefn {Target Hook} int TARGET_BRANCH_TARGET_REGISTER_CLASS (void)
9566 This target hook returns a register class for which branch target register
9567 optimizations should be applied. All registers in this class should be
9568 usable interchangeably. After reload, registers in this class will be
9569 re-allocated and loads will be hoisted out of loops and be subjected
9570 to inter-block scheduling.
9573 @deftypefn {Target Hook} bool TARGET_BRANCH_TARGET_REGISTER_CALLEE_SAVED (bool @var{after_prologue_epilogue_gen})
9574 Branch target register optimization will by default exclude callee-saved
9576 that are not already live during the current function; if this target hook
9577 returns true, they will be included. The target code must than make sure
9578 that all target registers in the class returned by
9579 @samp{TARGET_BRANCH_TARGET_REGISTER_CLASS} that might need saving are
9580 saved. @var{after_prologue_epilogue_gen} indicates if prologues and
9581 epilogues have already been generated. Note, even if you only return
9582 true when @var{after_prologue_epilogue_gen} is false, you still are likely
9583 to have to make special provisions in @code{INITIAL_ELIMINATION_OFFSET}
9584 to reserve space for caller-saved target registers.
9587 @defmac POWI_MAX_MULTS
9588 If defined, this macro is interpreted as a signed integer C expression
9589 that specifies the maximum number of floating point multiplications
9590 that should be emitted when expanding exponentiation by an integer
9591 constant inline. When this value is defined, exponentiation requiring
9592 more than this number of multiplications is implemented by calling the
9593 system library's @code{pow}, @code{powf} or @code{powl} routines.
9594 The default value places no upper bound on the multiplication count.
9597 @deftypefn Macro void TARGET_EXTRA_INCLUDES (const char *@var{sysroot}, const char *@var{iprefix}, int @var{stdinc})
9598 This target hook should register any extra include files for the
9599 target. The parameter @var{stdinc} indicates if normal include files
9600 are present. The parameter @var{sysroot} is the system root directory.
9601 The parameter @var{iprefix} is the prefix for the gcc directory.
9604 @deftypefn Macro void TARGET_EXTRA_PRE_INCLUDES (const char *@var{sysroot}, const char *@var{iprefix}, int @var{stdinc})
9605 This target hook should register any extra include files for the
9606 target before any standard headers. The parameter @var{stdinc}
9607 indicates if normal include files are present. The parameter
9608 @var{sysroot} is the system root directory. The parameter
9609 @var{iprefix} is the prefix for the gcc directory.
9612 @deftypefn Macro void TARGET_OPTF (char *@var{path})
9613 This target hook should register special include paths for the target.
9614 The parameter @var{path} is the include to register. On Darwin
9615 systems, this is used for Framework includes, which have semantics
9616 that are different from @option{-I}.
9619 @deftypefn {Target Hook} bool TARGET_USE_LOCAL_THUNK_ALIAS_P (tree @var{fndecl})
9620 This target hook returns @code{true} if it is safe to use a local alias
9621 for a virtual function @var{fndecl} when constructing thunks,
9622 @code{false} otherwise. By default, the hook returns @code{true} for all
9623 functions, if a target supports aliases (i.e.@: defines
9624 @code{ASM_OUTPUT_DEF}), @code{false} otherwise,
9627 @defmac TARGET_FORMAT_TYPES
9628 If defined, this macro is the name of a global variable containing
9629 target-specific format checking information for the @option{-Wformat}
9630 option. The default is to have no target-specific format checks.
9633 @defmac TARGET_N_FORMAT_TYPES
9634 If defined, this macro is the number of entries in
9635 @code{TARGET_FORMAT_TYPES}.
9638 @deftypefn {Target Hook} bool TARGET_RELAXED_ORDERING
9639 If set to @code{true}, means that the target's memory model does not
9640 guarantee that loads which do not depend on one another will access
9641 main memory in the order of the instruction stream; if ordering is
9642 important, an explicit memory barrier must be used. This is true of
9643 many recent processors which implement a policy of ``relaxed,''
9644 ``weak,'' or ``release'' memory consistency, such as Alpha, PowerPC,
9645 and ia64. The default is @code{false}.
9648 @deftypefn {Target Hook} const char *TARGET_INVALID_ARG_FOR_UNPROTOTYPED_FN (tree @var{typelist}, tree @var{funcdecl}, tree @var{val})
9649 If defined, this macro returns the diagnostic message when it is
9650 illegal to pass argument @var{val} to function @var{funcdecl}
9651 with prototype @var{typelist}.
9654 @deftypefn {Target Hook} {const char *} TARGET_INVALID_CONVERSION (tree @var{fromtype}, tree @var{totype})
9655 If defined, this macro returns the diagnostic message when it is
9656 invalid to convert from @var{fromtype} to @var{totype}, or @code{NULL}
9657 if validity should be determined by the front end.
9660 @deftypefn {Target Hook} {const char *} TARGET_INVALID_UNARY_OP (int @var{op}, tree @var{type})
9661 If defined, this macro returns the diagnostic message when it is
9662 invalid to apply operation @var{op} (where unary plus is denoted by
9663 @code{CONVERT_EXPR}) to an operand of type @var{type}, or @code{NULL}
9664 if validity should be determined by the front end.
9667 @deftypefn {Target Hook} {const char *} TARGET_INVALID_BINARY_OP (int @var{op}, tree @var{type1}, tree @var{type2})
9668 If defined, this macro returns the diagnostic message when it is
9669 invalid to apply operation @var{op} to operands of types @var{type1}
9670 and @var{type2}, or @code{NULL} if validity should be determined by
9674 @defmac TARGET_USE_JCR_SECTION
9675 This macro determines whether to use the JCR section to register Java
9676 classes. By default, TARGET_USE_JCR_SECTION is defined to 1 if both
9677 SUPPORTS_WEAK and TARGET_HAVE_NAMED_SECTIONS are true, else 0.