1 @c Copyright (C) 1988,1989,1992,1993,1994,1995,1996,1997,1998,1999,2000,2001,
2 @c 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010
3 @c Free Software Foundation, Inc.
4 @c This is part of the GCC manual.
5 @c For copying conditions, see the file gcc.texi.
8 @chapter Target Description Macros and Functions
9 @cindex machine description macros
10 @cindex target description macros
11 @cindex macros, target description
12 @cindex @file{tm.h} macros
14 In addition to the file @file{@var{machine}.md}, a machine description
15 includes a C header file conventionally given the name
16 @file{@var{machine}.h} and a C source file named @file{@var{machine}.c}.
17 The header file defines numerous macros that convey the information
18 about the target machine that does not fit into the scheme of the
19 @file{.md} file. The file @file{tm.h} should be a link to
20 @file{@var{machine}.h}. The header file @file{config.h} includes
21 @file{tm.h} and most compiler source files include @file{config.h}. The
22 source file defines a variable @code{targetm}, which is a structure
23 containing pointers to functions and data relating to the target
24 machine. @file{@var{machine}.c} should also contain their definitions,
25 if they are not defined elsewhere in GCC, and other functions called
26 through the macros defined in the @file{.h} file.
29 * Target Structure:: The @code{targetm} variable.
30 * Driver:: Controlling how the driver runs the compilation passes.
31 * Run-time Target:: Defining @samp{-m} options like @option{-m68000} and @option{-m68020}.
32 * Per-Function Data:: Defining data structures for per-function information.
33 * Storage Layout:: Defining sizes and alignments of data.
34 * Type Layout:: Defining sizes and properties of basic user data types.
35 * Registers:: Naming and describing the hardware registers.
36 * Register Classes:: Defining the classes of hardware registers.
37 * Old Constraints:: The old way to define machine-specific constraints.
38 * Stack and Calling:: Defining which way the stack grows and by how much.
39 * Varargs:: Defining the varargs macros.
40 * Trampolines:: Code set up at run time to enter a nested function.
41 * Library Calls:: Controlling how library routines are implicitly called.
42 * Addressing Modes:: Defining addressing modes valid for memory operands.
43 * Anchored Addresses:: Defining how @option{-fsection-anchors} should work.
44 * Condition Code:: Defining how insns update the condition code.
45 * Costs:: Defining relative costs of different operations.
46 * Scheduling:: Adjusting the behavior of the instruction scheduler.
47 * Sections:: Dividing storage into text, data, and other sections.
48 * PIC:: Macros for position independent code.
49 * Assembler Format:: Defining how to write insns and pseudo-ops to output.
50 * Debugging Info:: Defining the format of debugging output.
51 * Floating Point:: Handling floating point for cross-compilers.
52 * Mode Switching:: Insertion of mode-switching instructions.
53 * Target Attributes:: Defining target-specific uses of @code{__attribute__}.
54 * Emulated TLS:: Emulated TLS support.
55 * MIPS Coprocessors:: MIPS coprocessor support and how to customize it.
56 * PCH Target:: Validity checking for precompiled headers.
57 * C++ ABI:: Controlling C++ ABI changes.
58 * Named Address Spaces:: Adding support for named address spaces
59 * Misc:: Everything else.
62 @node Target Structure
63 @section The Global @code{targetm} Variable
65 @cindex target functions
67 @deftypevar {struct gcc_target} targetm
68 The target @file{.c} file must define the global @code{targetm} variable
69 which contains pointers to functions and data relating to the target
70 machine. The variable is declared in @file{target.h};
71 @file{target-def.h} defines the macro @code{TARGET_INITIALIZER} which is
72 used to initialize the variable, and macros for the default initializers
73 for elements of the structure. The @file{.c} file should override those
74 macros for which the default definition is inappropriate. For example:
77 #include "target-def.h"
79 /* @r{Initialize the GCC target structure.} */
81 #undef TARGET_COMP_TYPE_ATTRIBUTES
82 #define TARGET_COMP_TYPE_ATTRIBUTES @var{machine}_comp_type_attributes
84 struct gcc_target targetm = TARGET_INITIALIZER;
88 Where a macro should be defined in the @file{.c} file in this manner to
89 form part of the @code{targetm} structure, it is documented below as a
90 ``Target Hook'' with a prototype. Many macros will change in future
91 from being defined in the @file{.h} file to being part of the
92 @code{targetm} structure.
95 @section Controlling the Compilation Driver, @file{gcc}
97 @cindex controlling the compilation driver
99 @c prevent bad page break with this line
100 You can control the compilation driver.
102 @defmac SWITCH_TAKES_ARG (@var{char})
103 A C expression which determines whether the option @option{-@var{char}}
104 takes arguments. The value should be the number of arguments that
105 option takes--zero, for many options.
107 By default, this macro is defined as
108 @code{DEFAULT_SWITCH_TAKES_ARG}, which handles the standard options
109 properly. You need not define @code{SWITCH_TAKES_ARG} unless you
110 wish to add additional options which take arguments. Any redefinition
111 should call @code{DEFAULT_SWITCH_TAKES_ARG} and then check for
115 @defmac WORD_SWITCH_TAKES_ARG (@var{name})
116 A C expression which determines whether the option @option{-@var{name}}
117 takes arguments. The value should be the number of arguments that
118 option takes--zero, for many options. This macro rather than
119 @code{SWITCH_TAKES_ARG} is used for multi-character option names.
121 By default, this macro is defined as
122 @code{DEFAULT_WORD_SWITCH_TAKES_ARG}, which handles the standard options
123 properly. You need not define @code{WORD_SWITCH_TAKES_ARG} unless you
124 wish to add additional options which take arguments. Any redefinition
125 should call @code{DEFAULT_WORD_SWITCH_TAKES_ARG} and then check for
129 @defmac SWITCH_CURTAILS_COMPILATION (@var{char})
130 A C expression which determines whether the option @option{-@var{char}}
131 stops compilation before the generation of an executable. The value is
132 boolean, nonzero if the option does stop an executable from being
133 generated, zero otherwise.
135 By default, this macro is defined as
136 @code{DEFAULT_SWITCH_CURTAILS_COMPILATION}, which handles the standard
137 options properly. You need not define
138 @code{SWITCH_CURTAILS_COMPILATION} unless you wish to add additional
139 options which affect the generation of an executable. Any redefinition
140 should call @code{DEFAULT_SWITCH_CURTAILS_COMPILATION} and then check
141 for additional options.
144 @defmac TARGET_OPTION_TRANSLATE_TABLE
145 If defined, a list of pairs of strings, the first of which is a
146 potential command line target to the @file{gcc} driver program, and the
147 second of which is a space-separated (tabs and other whitespace are not
148 supported) list of options with which to replace the first option. The
149 target defining this list is responsible for assuring that the results
150 are valid. Replacement options may not be the @code{--opt} style, they
151 must be the @code{-opt} style. It is the intention of this macro to
152 provide a mechanism for substitution that affects the multilibs chosen,
153 such as one option that enables many options, some of which select
154 multilibs. Example nonsensical definition, where @option{-malt-abi},
155 @option{-EB}, and @option{-mspoo} cause different multilibs to be chosen:
158 #define TARGET_OPTION_TRANSLATE_TABLE \
159 @{ "-fast", "-march=fast-foo -malt-abi -I/usr/fast-foo" @}, \
160 @{ "-compat", "-EB -malign=4 -mspoo" @}
164 @defmac DRIVER_SELF_SPECS
165 A list of specs for the driver itself. It should be a suitable
166 initializer for an array of strings, with no surrounding braces.
168 The driver applies these specs to its own command line between loading
169 default @file{specs} files (but not command-line specified ones) and
170 choosing the multilib directory or running any subcommands. It
171 applies them in the order given, so each spec can depend on the
172 options added by earlier ones. It is also possible to remove options
173 using @samp{%<@var{option}} in the usual way.
175 This macro can be useful when a port has several interdependent target
176 options. It provides a way of standardizing the command line so
177 that the other specs are easier to write.
179 Do not define this macro if it does not need to do anything.
182 @defmac OPTION_DEFAULT_SPECS
183 A list of specs used to support configure-time default options (i.e.@:
184 @option{--with} options) in the driver. It should be a suitable initializer
185 for an array of structures, each containing two strings, without the
186 outermost pair of surrounding braces.
188 The first item in the pair is the name of the default. This must match
189 the code in @file{config.gcc} for the target. The second item is a spec
190 to apply if a default with this name was specified. The string
191 @samp{%(VALUE)} in the spec will be replaced by the value of the default
192 everywhere it occurs.
194 The driver will apply these specs to its own command line between loading
195 default @file{specs} files and processing @code{DRIVER_SELF_SPECS}, using
196 the same mechanism as @code{DRIVER_SELF_SPECS}.
198 Do not define this macro if it does not need to do anything.
202 A C string constant that tells the GCC driver program options to
203 pass to CPP@. It can also specify how to translate options you
204 give to GCC into options for GCC to pass to the CPP@.
206 Do not define this macro if it does not need to do anything.
209 @defmac CPLUSPLUS_CPP_SPEC
210 This macro is just like @code{CPP_SPEC}, but is used for C++, rather
211 than C@. If you do not define this macro, then the value of
212 @code{CPP_SPEC} (if any) will be used instead.
216 A C string constant that tells the GCC driver program options to
217 pass to @code{cc1}, @code{cc1plus}, @code{f771}, and the other language
219 It can also specify how to translate options you give to GCC into options
220 for GCC to pass to front ends.
222 Do not define this macro if it does not need to do anything.
226 A C string constant that tells the GCC driver program options to
227 pass to @code{cc1plus}. It can also specify how to translate options you
228 give to GCC into options for GCC to pass to the @code{cc1plus}.
230 Do not define this macro if it does not need to do anything.
231 Note that everything defined in CC1_SPEC is already passed to
232 @code{cc1plus} so there is no need to duplicate the contents of
233 CC1_SPEC in CC1PLUS_SPEC@.
237 A C string constant that tells the GCC driver program options to
238 pass to the assembler. It can also specify how to translate options
239 you give to GCC into options for GCC to pass to the assembler.
240 See the file @file{sun3.h} for an example of this.
242 Do not define this macro if it does not need to do anything.
245 @defmac ASM_FINAL_SPEC
246 A C string constant that tells the GCC driver program how to
247 run any programs which cleanup after the normal assembler.
248 Normally, this is not needed. See the file @file{mips.h} for
251 Do not define this macro if it does not need to do anything.
254 @defmac AS_NEEDS_DASH_FOR_PIPED_INPUT
255 Define this macro, with no value, if the driver should give the assembler
256 an argument consisting of a single dash, @option{-}, to instruct it to
257 read from its standard input (which will be a pipe connected to the
258 output of the compiler proper). This argument is given after any
259 @option{-o} option specifying the name of the output file.
261 If you do not define this macro, the assembler is assumed to read its
262 standard input if given no non-option arguments. If your assembler
263 cannot read standard input at all, use a @samp{%@{pipe:%e@}} construct;
264 see @file{mips.h} for instance.
268 A C string constant that tells the GCC driver program options to
269 pass to the linker. It can also specify how to translate options you
270 give to GCC into options for GCC to pass to the linker.
272 Do not define this macro if it does not need to do anything.
276 Another C string constant used much like @code{LINK_SPEC}. The difference
277 between the two is that @code{LIB_SPEC} is used at the end of the
278 command given to the linker.
280 If this macro is not defined, a default is provided that
281 loads the standard C library from the usual place. See @file{gcc.c}.
285 Another C string constant that tells the GCC driver program
286 how and when to place a reference to @file{libgcc.a} into the
287 linker command line. This constant is placed both before and after
288 the value of @code{LIB_SPEC}.
290 If this macro is not defined, the GCC driver provides a default that
291 passes the string @option{-lgcc} to the linker.
294 @defmac REAL_LIBGCC_SPEC
295 By default, if @code{ENABLE_SHARED_LIBGCC} is defined, the
296 @code{LIBGCC_SPEC} is not directly used by the driver program but is
297 instead modified to refer to different versions of @file{libgcc.a}
298 depending on the values of the command line flags @option{-static},
299 @option{-shared}, @option{-static-libgcc}, and @option{-shared-libgcc}. On
300 targets where these modifications are inappropriate, define
301 @code{REAL_LIBGCC_SPEC} instead. @code{REAL_LIBGCC_SPEC} tells the
302 driver how to place a reference to @file{libgcc} on the link command
303 line, but, unlike @code{LIBGCC_SPEC}, it is used unmodified.
306 @defmac USE_LD_AS_NEEDED
307 A macro that controls the modifications to @code{LIBGCC_SPEC}
308 mentioned in @code{REAL_LIBGCC_SPEC}. If nonzero, a spec will be
309 generated that uses --as-needed and the shared libgcc in place of the
310 static exception handler library, when linking without any of
311 @code{-static}, @code{-static-libgcc}, or @code{-shared-libgcc}.
315 If defined, this C string constant is added to @code{LINK_SPEC}.
316 When @code{USE_LD_AS_NEEDED} is zero or undefined, it also affects
317 the modifications to @code{LIBGCC_SPEC} mentioned in
318 @code{REAL_LIBGCC_SPEC}.
321 @defmac STARTFILE_SPEC
322 Another C string constant used much like @code{LINK_SPEC}. The
323 difference between the two is that @code{STARTFILE_SPEC} is used at
324 the very beginning of the command given to the linker.
326 If this macro is not defined, a default is provided that loads the
327 standard C startup file from the usual place. See @file{gcc.c}.
331 Another C string constant used much like @code{LINK_SPEC}. The
332 difference between the two is that @code{ENDFILE_SPEC} is used at
333 the very end of the command given to the linker.
335 Do not define this macro if it does not need to do anything.
338 @defmac THREAD_MODEL_SPEC
339 GCC @code{-v} will print the thread model GCC was configured to use.
340 However, this doesn't work on platforms that are multilibbed on thread
341 models, such as AIX 4.3. On such platforms, define
342 @code{THREAD_MODEL_SPEC} such that it evaluates to a string without
343 blanks that names one of the recognized thread models. @code{%*}, the
344 default value of this macro, will expand to the value of
345 @code{thread_file} set in @file{config.gcc}.
348 @defmac SYSROOT_SUFFIX_SPEC
349 Define this macro to add a suffix to the target sysroot when GCC is
350 configured with a sysroot. This will cause GCC to search for usr/lib,
351 et al, within sysroot+suffix.
354 @defmac SYSROOT_HEADERS_SUFFIX_SPEC
355 Define this macro to add a headers_suffix to the target sysroot when
356 GCC is configured with a sysroot. This will cause GCC to pass the
357 updated sysroot+headers_suffix to CPP, causing it to search for
358 usr/include, et al, within sysroot+headers_suffix.
362 Define this macro to provide additional specifications to put in the
363 @file{specs} file that can be used in various specifications like
366 The definition should be an initializer for an array of structures,
367 containing a string constant, that defines the specification name, and a
368 string constant that provides the specification.
370 Do not define this macro if it does not need to do anything.
372 @code{EXTRA_SPECS} is useful when an architecture contains several
373 related targets, which have various @code{@dots{}_SPECS} which are similar
374 to each other, and the maintainer would like one central place to keep
377 For example, the PowerPC System V.4 targets use @code{EXTRA_SPECS} to
378 define either @code{_CALL_SYSV} when the System V calling sequence is
379 used or @code{_CALL_AIX} when the older AIX-based calling sequence is
382 The @file{config/rs6000/rs6000.h} target file defines:
385 #define EXTRA_SPECS \
386 @{ "cpp_sysv_default", CPP_SYSV_DEFAULT @},
388 #define CPP_SYS_DEFAULT ""
391 The @file{config/rs6000/sysv.h} target file defines:
395 "%@{posix: -D_POSIX_SOURCE @} \
396 %@{mcall-sysv: -D_CALL_SYSV @} \
397 %@{!mcall-sysv: %(cpp_sysv_default) @} \
398 %@{msoft-float: -D_SOFT_FLOAT@} %@{mcpu=403: -D_SOFT_FLOAT@}"
400 #undef CPP_SYSV_DEFAULT
401 #define CPP_SYSV_DEFAULT "-D_CALL_SYSV"
404 while the @file{config/rs6000/eabiaix.h} target file defines
405 @code{CPP_SYSV_DEFAULT} as:
408 #undef CPP_SYSV_DEFAULT
409 #define CPP_SYSV_DEFAULT "-D_CALL_AIX"
413 @defmac LINK_LIBGCC_SPECIAL_1
414 Define this macro if the driver program should find the library
415 @file{libgcc.a}. If you do not define this macro, the driver program will pass
416 the argument @option{-lgcc} to tell the linker to do the search.
419 @defmac LINK_GCC_C_SEQUENCE_SPEC
420 The sequence in which libgcc and libc are specified to the linker.
421 By default this is @code{%G %L %G}.
424 @defmac LINK_COMMAND_SPEC
425 A C string constant giving the complete command line need to execute the
426 linker. When you do this, you will need to update your port each time a
427 change is made to the link command line within @file{gcc.c}. Therefore,
428 define this macro only if you need to completely redefine the command
429 line for invoking the linker and there is no other way to accomplish
430 the effect you need. Overriding this macro may be avoidable by overriding
431 @code{LINK_GCC_C_SEQUENCE_SPEC} instead.
434 @defmac LINK_ELIMINATE_DUPLICATE_LDIRECTORIES
435 A nonzero value causes @command{collect2} to remove duplicate @option{-L@var{directory}} search
436 directories from linking commands. Do not give it a nonzero value if
437 removing duplicate search directories changes the linker's semantics.
440 @defmac MULTILIB_DEFAULTS
441 Define this macro as a C expression for the initializer of an array of
442 string to tell the driver program which options are defaults for this
443 target and thus do not need to be handled specially when using
444 @code{MULTILIB_OPTIONS}.
446 Do not define this macro if @code{MULTILIB_OPTIONS} is not defined in
447 the target makefile fragment or if none of the options listed in
448 @code{MULTILIB_OPTIONS} are set by default.
449 @xref{Target Fragment}.
452 @defmac RELATIVE_PREFIX_NOT_LINKDIR
453 Define this macro to tell @command{gcc} that it should only translate
454 a @option{-B} prefix into a @option{-L} linker option if the prefix
455 indicates an absolute file name.
458 @defmac MD_EXEC_PREFIX
459 If defined, this macro is an additional prefix to try after
460 @code{STANDARD_EXEC_PREFIX}. @code{MD_EXEC_PREFIX} is not searched
461 when the compiler is built as a cross
462 compiler. If you define @code{MD_EXEC_PREFIX}, then be sure to add it
463 to the list of directories used to find the assembler in @file{configure.in}.
466 @defmac STANDARD_STARTFILE_PREFIX
467 Define this macro as a C string constant if you wish to override the
468 standard choice of @code{libdir} as the default prefix to
469 try when searching for startup files such as @file{crt0.o}.
470 @code{STANDARD_STARTFILE_PREFIX} is not searched when the compiler
471 is built as a cross compiler.
474 @defmac STANDARD_STARTFILE_PREFIX_1
475 Define this macro as a C string constant if you wish to override the
476 standard choice of @code{/lib} as a prefix to try after the default prefix
477 when searching for startup files such as @file{crt0.o}.
478 @code{STANDARD_STARTFILE_PREFIX_1} is not searched when the compiler
479 is built as a cross compiler.
482 @defmac STANDARD_STARTFILE_PREFIX_2
483 Define this macro as a C string constant if you wish to override the
484 standard choice of @code{/lib} as yet another prefix to try after the
485 default prefix when searching for startup files such as @file{crt0.o}.
486 @code{STANDARD_STARTFILE_PREFIX_2} is not searched when the compiler
487 is built as a cross compiler.
490 @defmac MD_STARTFILE_PREFIX
491 If defined, this macro supplies an additional prefix to try after the
492 standard prefixes. @code{MD_EXEC_PREFIX} is not searched when the
493 compiler is built as a cross compiler.
496 @defmac MD_STARTFILE_PREFIX_1
497 If defined, this macro supplies yet another prefix to try after the
498 standard prefixes. It is not searched when the compiler is built as a
502 @defmac INIT_ENVIRONMENT
503 Define this macro as a C string constant if you wish to set environment
504 variables for programs called by the driver, such as the assembler and
505 loader. The driver passes the value of this macro to @code{putenv} to
506 initialize the necessary environment variables.
509 @defmac LOCAL_INCLUDE_DIR
510 Define this macro as a C string constant if you wish to override the
511 standard choice of @file{/usr/local/include} as the default prefix to
512 try when searching for local header files. @code{LOCAL_INCLUDE_DIR}
513 comes before @code{SYSTEM_INCLUDE_DIR} in the search order.
515 Cross compilers do not search either @file{/usr/local/include} or its
519 @defmac SYSTEM_INCLUDE_DIR
520 Define this macro as a C string constant if you wish to specify a
521 system-specific directory to search for header files before the standard
522 directory. @code{SYSTEM_INCLUDE_DIR} comes before
523 @code{STANDARD_INCLUDE_DIR} in the search order.
525 Cross compilers do not use this macro and do not search the directory
529 @defmac STANDARD_INCLUDE_DIR
530 Define this macro as a C string constant if you wish to override the
531 standard choice of @file{/usr/include} as the default prefix to
532 try when searching for header files.
534 Cross compilers ignore this macro and do not search either
535 @file{/usr/include} or its replacement.
538 @defmac STANDARD_INCLUDE_COMPONENT
539 The ``component'' corresponding to @code{STANDARD_INCLUDE_DIR}.
540 See @code{INCLUDE_DEFAULTS}, below, for the description of components.
541 If you do not define this macro, no component is used.
544 @defmac INCLUDE_DEFAULTS
545 Define this macro if you wish to override the entire default search path
546 for include files. For a native compiler, the default search path
547 usually consists of @code{GCC_INCLUDE_DIR}, @code{LOCAL_INCLUDE_DIR},
548 @code{SYSTEM_INCLUDE_DIR}, @code{GPLUSPLUS_INCLUDE_DIR}, and
549 @code{STANDARD_INCLUDE_DIR}. In addition, @code{GPLUSPLUS_INCLUDE_DIR}
550 and @code{GCC_INCLUDE_DIR} are defined automatically by @file{Makefile},
551 and specify private search areas for GCC@. The directory
552 @code{GPLUSPLUS_INCLUDE_DIR} is used only for C++ programs.
554 The definition should be an initializer for an array of structures.
555 Each array element should have four elements: the directory name (a
556 string constant), the component name (also a string constant), a flag
557 for C++-only directories,
558 and a flag showing that the includes in the directory don't need to be
559 wrapped in @code{extern @samp{C}} when compiling C++. Mark the end of
560 the array with a null element.
562 The component name denotes what GNU package the include file is part of,
563 if any, in all uppercase letters. For example, it might be @samp{GCC}
564 or @samp{BINUTILS}. If the package is part of a vendor-supplied
565 operating system, code the component name as @samp{0}.
567 For example, here is the definition used for VAX/VMS:
570 #define INCLUDE_DEFAULTS \
572 @{ "GNU_GXX_INCLUDE:", "G++", 1, 1@}, \
573 @{ "GNU_CC_INCLUDE:", "GCC", 0, 0@}, \
574 @{ "SYS$SYSROOT:[SYSLIB.]", 0, 0, 0@}, \
581 Here is the order of prefixes tried for exec files:
585 Any prefixes specified by the user with @option{-B}.
588 The environment variable @code{GCC_EXEC_PREFIX} or, if @code{GCC_EXEC_PREFIX}
589 is not set and the compiler has not been installed in the configure-time
590 @var{prefix}, the location in which the compiler has actually been installed.
593 The directories specified by the environment variable @code{COMPILER_PATH}.
596 The macro @code{STANDARD_EXEC_PREFIX}, if the compiler has been installed
597 in the configured-time @var{prefix}.
600 The location @file{/usr/libexec/gcc/}, but only if this is a native compiler.
603 The location @file{/usr/lib/gcc/}, but only if this is a native compiler.
606 The macro @code{MD_EXEC_PREFIX}, if defined, but only if this is a native
610 Here is the order of prefixes tried for startfiles:
614 Any prefixes specified by the user with @option{-B}.
617 The environment variable @code{GCC_EXEC_PREFIX} or its automatically determined
618 value based on the installed toolchain location.
621 The directories specified by the environment variable @code{LIBRARY_PATH}
622 (or port-specific name; native only, cross compilers do not use this).
625 The macro @code{STANDARD_EXEC_PREFIX}, but only if the toolchain is installed
626 in the configured @var{prefix} or this is a native compiler.
629 The location @file{/usr/lib/gcc/}, but only if this is a native compiler.
632 The macro @code{MD_EXEC_PREFIX}, if defined, but only if this is a native
636 The macro @code{MD_STARTFILE_PREFIX}, if defined, but only if this is a
637 native compiler, or we have a target system root.
640 The macro @code{MD_STARTFILE_PREFIX_1}, if defined, but only if this is a
641 native compiler, or we have a target system root.
644 The macro @code{STANDARD_STARTFILE_PREFIX}, with any sysroot modifications.
645 If this path is relative it will be prefixed by @code{GCC_EXEC_PREFIX} and
646 the machine suffix or @code{STANDARD_EXEC_PREFIX} and the machine suffix.
649 The macro @code{STANDARD_STARTFILE_PREFIX_1}, but only if this is a native
650 compiler, or we have a target system root. The default for this macro is
654 The macro @code{STANDARD_STARTFILE_PREFIX_2}, but only if this is a native
655 compiler, or we have a target system root. The default for this macro is
659 @node Run-time Target
660 @section Run-time Target Specification
661 @cindex run-time target specification
662 @cindex predefined macros
663 @cindex target specifications
665 @c prevent bad page break with this line
666 Here are run-time target specifications.
668 @defmac TARGET_CPU_CPP_BUILTINS ()
669 This function-like macro expands to a block of code that defines
670 built-in preprocessor macros and assertions for the target CPU, using
671 the functions @code{builtin_define}, @code{builtin_define_std} and
672 @code{builtin_assert}. When the front end
673 calls this macro it provides a trailing semicolon, and since it has
674 finished command line option processing your code can use those
677 @code{builtin_assert} takes a string in the form you pass to the
678 command-line option @option{-A}, such as @code{cpu=mips}, and creates
679 the assertion. @code{builtin_define} takes a string in the form
680 accepted by option @option{-D} and unconditionally defines the macro.
682 @code{builtin_define_std} takes a string representing the name of an
683 object-like macro. If it doesn't lie in the user's namespace,
684 @code{builtin_define_std} defines it unconditionally. Otherwise, it
685 defines a version with two leading underscores, and another version
686 with two leading and trailing underscores, and defines the original
687 only if an ISO standard was not requested on the command line. For
688 example, passing @code{unix} defines @code{__unix}, @code{__unix__}
689 and possibly @code{unix}; passing @code{_mips} defines @code{__mips},
690 @code{__mips__} and possibly @code{_mips}, and passing @code{_ABI64}
691 defines only @code{_ABI64}.
693 You can also test for the C dialect being compiled. The variable
694 @code{c_language} is set to one of @code{clk_c}, @code{clk_cplusplus}
695 or @code{clk_objective_c}. Note that if we are preprocessing
696 assembler, this variable will be @code{clk_c} but the function-like
697 macro @code{preprocessing_asm_p()} will return true, so you might want
698 to check for that first. If you need to check for strict ANSI, the
699 variable @code{flag_iso} can be used. The function-like macro
700 @code{preprocessing_trad_p()} can be used to check for traditional
704 @defmac TARGET_OS_CPP_BUILTINS ()
705 Similarly to @code{TARGET_CPU_CPP_BUILTINS} but this macro is optional
706 and is used for the target operating system instead.
709 @defmac TARGET_OBJFMT_CPP_BUILTINS ()
710 Similarly to @code{TARGET_CPU_CPP_BUILTINS} but this macro is optional
711 and is used for the target object format. @file{elfos.h} uses this
712 macro to define @code{__ELF__}, so you probably do not need to define
716 @deftypevar {extern int} target_flags
717 This variable is declared in @file{options.h}, which is included before
718 any target-specific headers.
721 @deftypevr {Target Hook} int TARGET_DEFAULT_TARGET_FLAGS
722 This variable specifies the initial value of @code{target_flags}.
723 Its default setting is 0.
726 @cindex optional hardware or system features
727 @cindex features, optional, in system conventions
729 @deftypefn {Target Hook} bool TARGET_HANDLE_OPTION (size_t @var{code}, const char *@var{arg}, int @var{value})
730 This hook is called whenever the user specifies one of the
731 target-specific options described by the @file{.opt} definition files
732 (@pxref{Options}). It has the opportunity to do some option-specific
733 processing and should return true if the option is valid. The default
734 definition does nothing but return true.
736 @var{code} specifies the @code{OPT_@var{name}} enumeration value
737 associated with the selected option; @var{name} is just a rendering of
738 the option name in which non-alphanumeric characters are replaced by
739 underscores. @var{arg} specifies the string argument and is null if
740 no argument was given. If the option is flagged as a @code{UInteger}
741 (@pxref{Option properties}), @var{value} is the numeric value of the
742 argument. Otherwise @var{value} is 1 if the positive form of the
743 option was used and 0 if the ``no-'' form was.
746 @deftypefn {Target Hook} bool TARGET_HANDLE_C_OPTION (size_t @var{code}, const char *@var{arg}, int @var{value})
747 This target hook is called whenever the user specifies one of the
748 target-specific C language family options described by the @file{.opt}
749 definition files(@pxref{Options}). It has the opportunity to do some
750 option-specific processing and should return true if the option is
751 valid. The arguments are like for @code{TARGET_HANDLE_OPTION}. The
752 default definition does nothing but return false.
754 In general, you should use @code{TARGET_HANDLE_OPTION} to handle
755 options. However, if processing an option requires routines that are
756 only available in the C (and related language) front ends, then you
757 should use @code{TARGET_HANDLE_C_OPTION} instead.
760 @defmac TARGET_VERSION
761 This macro is a C statement to print on @code{stderr} a string
762 describing the particular machine description choice. Every machine
763 description should define @code{TARGET_VERSION}. For example:
767 #define TARGET_VERSION \
768 fprintf (stderr, " (68k, Motorola syntax)");
770 #define TARGET_VERSION \
771 fprintf (stderr, " (68k, MIT syntax)");
776 @deftypefn {Target Hook} void TARGET_OVERRIDE_OPTIONS_AFTER_CHANGE (void)
777 This target function is similar to the hook @code{TARGET_OPTION_OVERRIDE}
778 but is called when the optimize level is changed via an attribute or
779 pragma or when it is reset at the end of the code affected by the
780 attribute or pragma. It is not called at the beginning of compilation
781 when @code{TARGET_OPTION_OVERRIDE} is called so if you want to perform these
782 actions then, you should have @code{TARGET_OPTION_OVERRIDE} call
783 @code{TARGET_OVERRIDE_OPTIONS_AFTER_CHANGE}.
786 @defmac C_COMMON_OVERRIDE_OPTIONS
787 This is similar to the @code{TARGET_OPTION_OVERRIDE} hook
788 but is only used in the C
789 language frontends (C, Objective-C, C++, Objective-C++) and so can be
790 used to alter option flag variables which only exist in those
794 @defmac OPTIMIZATION_OPTIONS (@var{level}, @var{size})
795 Some machines may desire to change what optimizations are performed for
796 various optimization levels. This macro, if defined, is executed once
797 just after the optimization level is determined and before the remainder
798 of the command options have been parsed. Values set in this macro are
799 used as the default values for the other command line options.
801 @var{level} is the optimization level specified; 2 if @option{-O2} is
802 specified, 1 if @option{-O} is specified, and 0 if neither is specified.
804 @var{size} is nonzero if @option{-Os} is specified and zero otherwise.
806 This macro is run once at program startup and when the optimization
807 options are changed via @code{#pragma GCC optimize} or by using the
808 @code{optimize} attribute.
810 @strong{Do not examine @code{write_symbols} in
811 this macro!} The debugging options are not supposed to alter the
815 @deftypefn {Target Hook} void TARGET_HELP (void)
816 This hook is called in response to the user invoking
817 @option{--target-help} on the command line. It gives the target a
818 chance to display extra information on the target specific command
819 line options found in its @file{.opt} file.
822 @defmac CAN_DEBUG_WITHOUT_FP
823 Define this macro if debugging can be performed even without a frame
824 pointer. If this macro is defined, GCC will turn on the
825 @option{-fomit-frame-pointer} option whenever @option{-O} is specified.
828 @defmac SWITCHABLE_TARGET
829 Some targets need to switch between substantially different subtargets
830 during compilation. For example, the MIPS target has one subtarget for
831 the traditional MIPS architecture and another for MIPS16. Source code
832 can switch between these two subarchitectures using the @code{mips16}
833 and @code{nomips16} attributes.
835 Such subtargets can differ in things like the set of available
836 registers, the set of available instructions, the costs of various
837 operations, and so on. GCC caches a lot of this type of information
838 in global variables, and recomputing them for each subtarget takes a
839 significant amount of time. The compiler therefore provides a facility
840 for maintaining several versions of the global variables and quickly
841 switching between them; see @file{target-globals.h} for details.
843 Define this macro to 1 if your target needs this facility. The default
847 @node Per-Function Data
848 @section Defining data structures for per-function information.
849 @cindex per-function data
850 @cindex data structures
852 If the target needs to store information on a per-function basis, GCC
853 provides a macro and a couple of variables to allow this. Note, just
854 using statics to store the information is a bad idea, since GCC supports
855 nested functions, so you can be halfway through encoding one function
856 when another one comes along.
858 GCC defines a data structure called @code{struct function} which
859 contains all of the data specific to an individual function. This
860 structure contains a field called @code{machine} whose type is
861 @code{struct machine_function *}, which can be used by targets to point
862 to their own specific data.
864 If a target needs per-function specific data it should define the type
865 @code{struct machine_function} and also the macro @code{INIT_EXPANDERS}.
866 This macro should be used to initialize the function pointer
867 @code{init_machine_status}. This pointer is explained below.
869 One typical use of per-function, target specific data is to create an
870 RTX to hold the register containing the function's return address. This
871 RTX can then be used to implement the @code{__builtin_return_address}
872 function, for level 0.
874 Note---earlier implementations of GCC used a single data area to hold
875 all of the per-function information. Thus when processing of a nested
876 function began the old per-function data had to be pushed onto a
877 stack, and when the processing was finished, it had to be popped off the
878 stack. GCC used to provide function pointers called
879 @code{save_machine_status} and @code{restore_machine_status} to handle
880 the saving and restoring of the target specific information. Since the
881 single data area approach is no longer used, these pointers are no
884 @defmac INIT_EXPANDERS
885 Macro called to initialize any target specific information. This macro
886 is called once per function, before generation of any RTL has begun.
887 The intention of this macro is to allow the initialization of the
888 function pointer @code{init_machine_status}.
891 @deftypevar {void (*)(struct function *)} init_machine_status
892 If this function pointer is non-@code{NULL} it will be called once per
893 function, before function compilation starts, in order to allow the
894 target to perform any target specific initialization of the
895 @code{struct function} structure. It is intended that this would be
896 used to initialize the @code{machine} of that structure.
898 @code{struct machine_function} structures are expected to be freed by GC@.
899 Generally, any memory that they reference must be allocated by using
900 GC allocation, including the structure itself.
904 @section Storage Layout
905 @cindex storage layout
907 Note that the definitions of the macros in this table which are sizes or
908 alignments measured in bits do not need to be constant. They can be C
909 expressions that refer to static variables, such as the @code{target_flags}.
910 @xref{Run-time Target}.
912 @defmac BITS_BIG_ENDIAN
913 Define this macro to have the value 1 if the most significant bit in a
914 byte has the lowest number; otherwise define it to have the value zero.
915 This means that bit-field instructions count from the most significant
916 bit. If the machine has no bit-field instructions, then this must still
917 be defined, but it doesn't matter which value it is defined to. This
918 macro need not be a constant.
920 This macro does not affect the way structure fields are packed into
921 bytes or words; that is controlled by @code{BYTES_BIG_ENDIAN}.
924 @defmac BYTES_BIG_ENDIAN
925 Define this macro to have the value 1 if the most significant byte in a
926 word has the lowest number. This macro need not be a constant.
929 @defmac WORDS_BIG_ENDIAN
930 Define this macro to have the value 1 if, in a multiword object, the
931 most significant word has the lowest number. This applies to both
932 memory locations and registers; GCC fundamentally assumes that the
933 order of words in memory is the same as the order in registers. This
934 macro need not be a constant.
937 @defmac LIBGCC2_WORDS_BIG_ENDIAN
938 Define this macro if @code{WORDS_BIG_ENDIAN} is not constant. This must be a
939 constant value with the same meaning as @code{WORDS_BIG_ENDIAN}, which will be
940 used only when compiling @file{libgcc2.c}. Typically the value will be set
941 based on preprocessor defines.
944 @defmac FLOAT_WORDS_BIG_ENDIAN
945 Define this macro to have the value 1 if @code{DFmode}, @code{XFmode} or
946 @code{TFmode} floating point numbers are stored in memory with the word
947 containing the sign bit at the lowest address; otherwise define it to
948 have the value 0. This macro need not be a constant.
950 You need not define this macro if the ordering is the same as for
954 @defmac BITS_PER_UNIT
955 Define this macro to be the number of bits in an addressable storage
956 unit (byte). If you do not define this macro the default is 8.
959 @defmac BITS_PER_WORD
960 Number of bits in a word. If you do not define this macro, the default
961 is @code{BITS_PER_UNIT * UNITS_PER_WORD}.
964 @defmac MAX_BITS_PER_WORD
965 Maximum number of bits in a word. If this is undefined, the default is
966 @code{BITS_PER_WORD}. Otherwise, it is the constant value that is the
967 largest value that @code{BITS_PER_WORD} can have at run-time.
970 @defmac UNITS_PER_WORD
971 Number of storage units in a word; normally the size of a general-purpose
972 register, a power of two from 1 or 8.
975 @defmac MIN_UNITS_PER_WORD
976 Minimum number of units in a word. If this is undefined, the default is
977 @code{UNITS_PER_WORD}. Otherwise, it is the constant value that is the
978 smallest value that @code{UNITS_PER_WORD} can have at run-time.
982 Width of a pointer, in bits. You must specify a value no wider than the
983 width of @code{Pmode}. If it is not equal to the width of @code{Pmode},
984 you must define @code{POINTERS_EXTEND_UNSIGNED}. If you do not specify
985 a value the default is @code{BITS_PER_WORD}.
988 @defmac POINTERS_EXTEND_UNSIGNED
989 A C expression that determines how pointers should be extended from
990 @code{ptr_mode} to either @code{Pmode} or @code{word_mode}. It is
991 greater than zero if pointers should be zero-extended, zero if they
992 should be sign-extended, and negative if some other sort of conversion
993 is needed. In the last case, the extension is done by the target's
994 @code{ptr_extend} instruction.
996 You need not define this macro if the @code{ptr_mode}, @code{Pmode}
997 and @code{word_mode} are all the same width.
1000 @defmac PROMOTE_MODE (@var{m}, @var{unsignedp}, @var{type})
1001 A macro to update @var{m} and @var{unsignedp} when an object whose type
1002 is @var{type} and which has the specified mode and signedness is to be
1003 stored in a register. This macro is only called when @var{type} is a
1006 On most RISC machines, which only have operations that operate on a full
1007 register, define this macro to set @var{m} to @code{word_mode} if
1008 @var{m} is an integer mode narrower than @code{BITS_PER_WORD}. In most
1009 cases, only integer modes should be widened because wider-precision
1010 floating-point operations are usually more expensive than their narrower
1013 For most machines, the macro definition does not change @var{unsignedp}.
1014 However, some machines, have instructions that preferentially handle
1015 either signed or unsigned quantities of certain modes. For example, on
1016 the DEC Alpha, 32-bit loads from memory and 32-bit add instructions
1017 sign-extend the result to 64 bits. On such machines, set
1018 @var{unsignedp} according to which kind of extension is more efficient.
1020 Do not define this macro if it would never modify @var{m}.
1023 @deftypefn {Target Hook} {enum machine_mode} TARGET_PROMOTE_FUNCTION_MODE (const_tree @var{type}, enum machine_mode @var{mode}, int *@var{punsignedp}, const_tree @var{funtype}, int @var{for_return})
1024 Like @code{PROMOTE_MODE}, but it is applied to outgoing function arguments or
1025 function return values. The target hook should return the new mode
1026 and possibly change @code{*@var{punsignedp}} if the promotion should
1027 change signedness. This function is called only for scalar @emph{or
1030 @var{for_return} allows to distinguish the promotion of arguments and
1031 return values. If it is @code{1}, a return value is being promoted and
1032 @code{TARGET_FUNCTION_VALUE} must perform the same promotions done here.
1033 If it is @code{2}, the returned mode should be that of the register in
1034 which an incoming parameter is copied, or the outgoing result is computed;
1035 then the hook should return the same mode as @code{promote_mode}, though
1036 the signedness may be different.
1038 The default is to not promote arguments and return values. You can
1039 also define the hook to @code{default_promote_function_mode_always_promote}
1040 if you would like to apply the same rules given by @code{PROMOTE_MODE}.
1043 @defmac PARM_BOUNDARY
1044 Normal alignment required for function parameters on the stack, in
1045 bits. All stack parameters receive at least this much alignment
1046 regardless of data type. On most machines, this is the same as the
1050 @defmac STACK_BOUNDARY
1051 Define this macro to the minimum alignment enforced by hardware for the
1052 stack pointer on this machine. The definition is a C expression for the
1053 desired alignment (measured in bits). This value is used as a default
1054 if @code{PREFERRED_STACK_BOUNDARY} is not defined. On most machines,
1055 this should be the same as @code{PARM_BOUNDARY}.
1058 @defmac PREFERRED_STACK_BOUNDARY
1059 Define this macro if you wish to preserve a certain alignment for the
1060 stack pointer, greater than what the hardware enforces. The definition
1061 is a C expression for the desired alignment (measured in bits). This
1062 macro must evaluate to a value equal to or larger than
1063 @code{STACK_BOUNDARY}.
1066 @defmac INCOMING_STACK_BOUNDARY
1067 Define this macro if the incoming stack boundary may be different
1068 from @code{PREFERRED_STACK_BOUNDARY}. This macro must evaluate
1069 to a value equal to or larger than @code{STACK_BOUNDARY}.
1072 @defmac FUNCTION_BOUNDARY
1073 Alignment required for a function entry point, in bits.
1076 @defmac BIGGEST_ALIGNMENT
1077 Biggest alignment that any data type can require on this machine, in
1078 bits. Note that this is not the biggest alignment that is supported,
1079 just the biggest alignment that, when violated, may cause a fault.
1082 @defmac MALLOC_ABI_ALIGNMENT
1083 Alignment, in bits, a C conformant malloc implementation has to
1084 provide. If not defined, the default value is @code{BITS_PER_WORD}.
1087 @defmac ATTRIBUTE_ALIGNED_VALUE
1088 Alignment used by the @code{__attribute__ ((aligned))} construct. If
1089 not defined, the default value is @code{BIGGEST_ALIGNMENT}.
1092 @defmac MINIMUM_ATOMIC_ALIGNMENT
1093 If defined, the smallest alignment, in bits, that can be given to an
1094 object that can be referenced in one operation, without disturbing any
1095 nearby object. Normally, this is @code{BITS_PER_UNIT}, but may be larger
1096 on machines that don't have byte or half-word store operations.
1099 @defmac BIGGEST_FIELD_ALIGNMENT
1100 Biggest alignment that any structure or union field can require on this
1101 machine, in bits. If defined, this overrides @code{BIGGEST_ALIGNMENT} for
1102 structure and union fields only, unless the field alignment has been set
1103 by the @code{__attribute__ ((aligned (@var{n})))} construct.
1106 @defmac ADJUST_FIELD_ALIGN (@var{field}, @var{computed})
1107 An expression for the alignment of a structure field @var{field} if the
1108 alignment computed in the usual way (including applying of
1109 @code{BIGGEST_ALIGNMENT} and @code{BIGGEST_FIELD_ALIGNMENT} to the
1110 alignment) is @var{computed}. It overrides alignment only if the
1111 field alignment has not been set by the
1112 @code{__attribute__ ((aligned (@var{n})))} construct.
1115 @defmac MAX_STACK_ALIGNMENT
1116 Biggest stack alignment guaranteed by the backend. Use this macro
1117 to specify the maximum alignment of a variable on stack.
1119 If not defined, the default value is @code{STACK_BOUNDARY}.
1121 @c FIXME: The default should be @code{PREFERRED_STACK_BOUNDARY}.
1122 @c But the fix for PR 32893 indicates that we can only guarantee
1123 @c maximum stack alignment on stack up to @code{STACK_BOUNDARY}, not
1124 @c @code{PREFERRED_STACK_BOUNDARY}, if stack alignment isn't supported.
1127 @defmac MAX_OFILE_ALIGNMENT
1128 Biggest alignment supported by the object file format of this machine.
1129 Use this macro to limit the alignment which can be specified using the
1130 @code{__attribute__ ((aligned (@var{n})))} construct. If not defined,
1131 the default value is @code{BIGGEST_ALIGNMENT}.
1133 On systems that use ELF, the default (in @file{config/elfos.h}) is
1134 the largest supported 32-bit ELF section alignment representable on
1135 a 32-bit host e.g. @samp{(((unsigned HOST_WIDEST_INT) 1 << 28) * 8)}.
1136 On 32-bit ELF the largest supported section alignment in bits is
1137 @samp{(0x80000000 * 8)}, but this is not representable on 32-bit hosts.
1140 @defmac DATA_ALIGNMENT (@var{type}, @var{basic-align})
1141 If defined, a C expression to compute the alignment for a variable in
1142 the static store. @var{type} is the data type, and @var{basic-align} is
1143 the alignment that the object would ordinarily have. The value of this
1144 macro is used instead of that alignment to align the object.
1146 If this macro is not defined, then @var{basic-align} is used.
1149 One use of this macro is to increase alignment of medium-size data to
1150 make it all fit in fewer cache lines. Another is to cause character
1151 arrays to be word-aligned so that @code{strcpy} calls that copy
1152 constants to character arrays can be done inline.
1155 @defmac CONSTANT_ALIGNMENT (@var{constant}, @var{basic-align})
1156 If defined, a C expression to compute the alignment given to a constant
1157 that is being placed in memory. @var{constant} is the constant and
1158 @var{basic-align} is the alignment that the object would ordinarily
1159 have. The value of this macro is used instead of that alignment to
1162 If this macro is not defined, then @var{basic-align} is used.
1164 The typical use of this macro is to increase alignment for string
1165 constants to be word aligned so that @code{strcpy} calls that copy
1166 constants can be done inline.
1169 @defmac LOCAL_ALIGNMENT (@var{type}, @var{basic-align})
1170 If defined, a C expression to compute the alignment for a variable in
1171 the local store. @var{type} is the data type, and @var{basic-align} is
1172 the alignment that the object would ordinarily have. The value of this
1173 macro is used instead of that alignment to align the object.
1175 If this macro is not defined, then @var{basic-align} is used.
1177 One use of this macro is to increase alignment of medium-size data to
1178 make it all fit in fewer cache lines.
1181 @defmac STACK_SLOT_ALIGNMENT (@var{type}, @var{mode}, @var{basic-align})
1182 If defined, a C expression to compute the alignment for stack slot.
1183 @var{type} is the data type, @var{mode} is the widest mode available,
1184 and @var{basic-align} is the alignment that the slot would ordinarily
1185 have. The value of this macro is used instead of that alignment to
1188 If this macro is not defined, then @var{basic-align} is used when
1189 @var{type} is @code{NULL}. Otherwise, @code{LOCAL_ALIGNMENT} will
1192 This macro is to set alignment of stack slot to the maximum alignment
1193 of all possible modes which the slot may have.
1196 @defmac LOCAL_DECL_ALIGNMENT (@var{decl})
1197 If defined, a C expression to compute the alignment for a local
1198 variable @var{decl}.
1200 If this macro is not defined, then
1201 @code{LOCAL_ALIGNMENT (TREE_TYPE (@var{decl}), DECL_ALIGN (@var{decl}))}
1204 One use of this macro is to increase alignment of medium-size data to
1205 make it all fit in fewer cache lines.
1208 @defmac MINIMUM_ALIGNMENT (@var{exp}, @var{mode}, @var{align})
1209 If defined, a C expression to compute the minimum required alignment
1210 for dynamic stack realignment purposes for @var{exp} (a type or decl),
1211 @var{mode}, assuming normal alignment @var{align}.
1213 If this macro is not defined, then @var{align} will be used.
1216 @defmac EMPTY_FIELD_BOUNDARY
1217 Alignment in bits to be given to a structure bit-field that follows an
1218 empty field such as @code{int : 0;}.
1220 If @code{PCC_BITFIELD_TYPE_MATTERS} is true, it overrides this macro.
1223 @defmac STRUCTURE_SIZE_BOUNDARY
1224 Number of bits which any structure or union's size must be a multiple of.
1225 Each structure or union's size is rounded up to a multiple of this.
1227 If you do not define this macro, the default is the same as
1228 @code{BITS_PER_UNIT}.
1231 @defmac STRICT_ALIGNMENT
1232 Define this macro to be the value 1 if instructions will fail to work
1233 if given data not on the nominal alignment. If instructions will merely
1234 go slower in that case, define this macro as 0.
1237 @defmac PCC_BITFIELD_TYPE_MATTERS
1238 Define this if you wish to imitate the way many other C compilers handle
1239 alignment of bit-fields and the structures that contain them.
1241 The behavior is that the type written for a named bit-field (@code{int},
1242 @code{short}, or other integer type) imposes an alignment for the entire
1243 structure, as if the structure really did contain an ordinary field of
1244 that type. In addition, the bit-field is placed within the structure so
1245 that it would fit within such a field, not crossing a boundary for it.
1247 Thus, on most machines, a named bit-field whose type is written as
1248 @code{int} would not cross a four-byte boundary, and would force
1249 four-byte alignment for the whole structure. (The alignment used may
1250 not be four bytes; it is controlled by the other alignment parameters.)
1252 An unnamed bit-field will not affect the alignment of the containing
1255 If the macro is defined, its definition should be a C expression;
1256 a nonzero value for the expression enables this behavior.
1258 Note that if this macro is not defined, or its value is zero, some
1259 bit-fields may cross more than one alignment boundary. The compiler can
1260 support such references if there are @samp{insv}, @samp{extv}, and
1261 @samp{extzv} insns that can directly reference memory.
1263 The other known way of making bit-fields work is to define
1264 @code{STRUCTURE_SIZE_BOUNDARY} as large as @code{BIGGEST_ALIGNMENT}.
1265 Then every structure can be accessed with fullwords.
1267 Unless the machine has bit-field instructions or you define
1268 @code{STRUCTURE_SIZE_BOUNDARY} that way, you must define
1269 @code{PCC_BITFIELD_TYPE_MATTERS} to have a nonzero value.
1271 If your aim is to make GCC use the same conventions for laying out
1272 bit-fields as are used by another compiler, here is how to investigate
1273 what the other compiler does. Compile and run this program:
1292 printf ("Size of foo1 is %d\n",
1293 sizeof (struct foo1));
1294 printf ("Size of foo2 is %d\n",
1295 sizeof (struct foo2));
1300 If this prints 2 and 5, then the compiler's behavior is what you would
1301 get from @code{PCC_BITFIELD_TYPE_MATTERS}.
1304 @defmac BITFIELD_NBYTES_LIMITED
1305 Like @code{PCC_BITFIELD_TYPE_MATTERS} except that its effect is limited
1306 to aligning a bit-field within the structure.
1309 @deftypefn {Target Hook} bool TARGET_ALIGN_ANON_BITFIELD (void)
1310 When @code{PCC_BITFIELD_TYPE_MATTERS} is true this hook will determine
1311 whether unnamed bitfields affect the alignment of the containing
1312 structure. The hook should return true if the structure should inherit
1313 the alignment requirements of an unnamed bitfield's type.
1316 @deftypefn {Target Hook} bool TARGET_NARROW_VOLATILE_BITFIELD (void)
1317 This target hook should return @code{true} if accesses to volatile bitfields
1318 should use the narrowest mode possible. It should return @code{false} if
1319 these accesses should use the bitfield container type.
1321 The default is @code{!TARGET_STRICT_ALIGN}.
1324 @defmac MEMBER_TYPE_FORCES_BLK (@var{field}, @var{mode})
1325 Return 1 if a structure or array containing @var{field} should be accessed using
1328 If @var{field} is the only field in the structure, @var{mode} is its
1329 mode, otherwise @var{mode} is VOIDmode. @var{mode} is provided in the
1330 case where structures of one field would require the structure's mode to
1331 retain the field's mode.
1333 Normally, this is not needed.
1336 @defmac ROUND_TYPE_ALIGN (@var{type}, @var{computed}, @var{specified})
1337 Define this macro as an expression for the alignment of a type (given
1338 by @var{type} as a tree node) if the alignment computed in the usual
1339 way is @var{computed} and the alignment explicitly specified was
1342 The default is to use @var{specified} if it is larger; otherwise, use
1343 the smaller of @var{computed} and @code{BIGGEST_ALIGNMENT}
1346 @defmac MAX_FIXED_MODE_SIZE
1347 An integer expression for the size in bits of the largest integer
1348 machine mode that should actually be used. All integer machine modes of
1349 this size or smaller can be used for structures and unions with the
1350 appropriate sizes. If this macro is undefined, @code{GET_MODE_BITSIZE
1351 (DImode)} is assumed.
1354 @defmac STACK_SAVEAREA_MODE (@var{save_level})
1355 If defined, an expression of type @code{enum machine_mode} that
1356 specifies the mode of the save area operand of a
1357 @code{save_stack_@var{level}} named pattern (@pxref{Standard Names}).
1358 @var{save_level} is one of @code{SAVE_BLOCK}, @code{SAVE_FUNCTION}, or
1359 @code{SAVE_NONLOCAL} and selects which of the three named patterns is
1360 having its mode specified.
1362 You need not define this macro if it always returns @code{Pmode}. You
1363 would most commonly define this macro if the
1364 @code{save_stack_@var{level}} patterns need to support both a 32- and a
1368 @defmac STACK_SIZE_MODE
1369 If defined, an expression of type @code{enum machine_mode} that
1370 specifies the mode of the size increment operand of an
1371 @code{allocate_stack} named pattern (@pxref{Standard Names}).
1373 You need not define this macro if it always returns @code{word_mode}.
1374 You would most commonly define this macro if the @code{allocate_stack}
1375 pattern needs to support both a 32- and a 64-bit mode.
1378 @deftypefn {Target Hook} {enum machine_mode} TARGET_LIBGCC_CMP_RETURN_MODE (void)
1379 This target hook should return the mode to be used for the return value
1380 of compare instructions expanded to libgcc calls. If not defined
1381 @code{word_mode} is returned which is the right choice for a majority of
1385 @deftypefn {Target Hook} {enum machine_mode} TARGET_LIBGCC_SHIFT_COUNT_MODE (void)
1386 This target hook should return the mode to be used for the shift count operand
1387 of shift instructions expanded to libgcc calls. If not defined
1388 @code{word_mode} is returned which is the right choice for a majority of
1392 @deftypefn {Target Hook} {enum machine_mode} TARGET_UNWIND_WORD_MODE (void)
1393 Return machine mode to be used for @code{_Unwind_Word} type.
1394 The default is to use @code{word_mode}.
1397 @defmac ROUND_TOWARDS_ZERO
1398 If defined, this macro should be true if the prevailing rounding
1399 mode is towards zero.
1401 Defining this macro only affects the way @file{libgcc.a} emulates
1402 floating-point arithmetic.
1404 Not defining this macro is equivalent to returning zero.
1407 @defmac LARGEST_EXPONENT_IS_NORMAL (@var{size})
1408 This macro should return true if floats with @var{size}
1409 bits do not have a NaN or infinity representation, but use the largest
1410 exponent for normal numbers instead.
1412 Defining this macro only affects the way @file{libgcc.a} emulates
1413 floating-point arithmetic.
1415 The default definition of this macro returns false for all sizes.
1418 @deftypefn {Target Hook} bool TARGET_MS_BITFIELD_LAYOUT_P (const_tree @var{record_type})
1419 This target hook returns @code{true} if bit-fields in the given
1420 @var{record_type} are to be laid out following the rules of Microsoft
1421 Visual C/C++, namely: (i) a bit-field won't share the same storage
1422 unit with the previous bit-field if their underlying types have
1423 different sizes, and the bit-field will be aligned to the highest
1424 alignment of the underlying types of itself and of the previous
1425 bit-field; (ii) a zero-sized bit-field will affect the alignment of
1426 the whole enclosing structure, even if it is unnamed; except that
1427 (iii) a zero-sized bit-field will be disregarded unless it follows
1428 another bit-field of nonzero size. If this hook returns @code{true},
1429 other macros that control bit-field layout are ignored.
1431 When a bit-field is inserted into a packed record, the whole size
1432 of the underlying type is used by one or more same-size adjacent
1433 bit-fields (that is, if its long:3, 32 bits is used in the record,
1434 and any additional adjacent long bit-fields are packed into the same
1435 chunk of 32 bits. However, if the size changes, a new field of that
1436 size is allocated). In an unpacked record, this is the same as using
1437 alignment, but not equivalent when packing.
1439 If both MS bit-fields and @samp{__attribute__((packed))} are used,
1440 the latter will take precedence. If @samp{__attribute__((packed))} is
1441 used on a single field when MS bit-fields are in use, it will take
1442 precedence for that field, but the alignment of the rest of the structure
1443 may affect its placement.
1446 @deftypefn {Target Hook} bool TARGET_DECIMAL_FLOAT_SUPPORTED_P (void)
1447 Returns true if the target supports decimal floating point.
1450 @deftypefn {Target Hook} bool TARGET_FIXED_POINT_SUPPORTED_P (void)
1451 Returns true if the target supports fixed-point arithmetic.
1454 @deftypefn {Target Hook} void TARGET_EXPAND_TO_RTL_HOOK (void)
1455 This hook is called just before expansion into rtl, allowing the target
1456 to perform additional initializations or analysis before the expansion.
1457 For example, the rs6000 port uses it to allocate a scratch stack slot
1458 for use in copying SDmode values between memory and floating point
1459 registers whenever the function being expanded has any SDmode
1463 @deftypefn {Target Hook} void TARGET_INSTANTIATE_DECLS (void)
1464 This hook allows the backend to perform additional instantiations on rtl
1465 that are not actually in any insns yet, but will be later.
1468 @deftypefn {Target Hook} {const char *} TARGET_MANGLE_TYPE (const_tree @var{type})
1469 If your target defines any fundamental types, or any types your target
1470 uses should be mangled differently from the default, define this hook
1471 to return the appropriate encoding for these types as part of a C++
1472 mangled name. The @var{type} argument is the tree structure representing
1473 the type to be mangled. The hook may be applied to trees which are
1474 not target-specific fundamental types; it should return @code{NULL}
1475 for all such types, as well as arguments it does not recognize. If the
1476 return value is not @code{NULL}, it must point to a statically-allocated
1479 Target-specific fundamental types might be new fundamental types or
1480 qualified versions of ordinary fundamental types. Encode new
1481 fundamental types as @samp{@w{u @var{n} @var{name}}}, where @var{name}
1482 is the name used for the type in source code, and @var{n} is the
1483 length of @var{name} in decimal. Encode qualified versions of
1484 ordinary types as @samp{@w{U @var{n} @var{name} @var{code}}}, where
1485 @var{name} is the name used for the type qualifier in source code,
1486 @var{n} is the length of @var{name} as above, and @var{code} is the
1487 code used to represent the unqualified version of this type. (See
1488 @code{write_builtin_type} in @file{cp/mangle.c} for the list of
1489 codes.) In both cases the spaces are for clarity; do not include any
1490 spaces in your string.
1492 This hook is applied to types prior to typedef resolution. If the mangled
1493 name for a particular type depends only on that type's main variant, you
1494 can perform typedef resolution yourself using @code{TYPE_MAIN_VARIANT}
1497 The default version of this hook always returns @code{NULL}, which is
1498 appropriate for a target that does not define any new fundamental
1503 @section Layout of Source Language Data Types
1505 These macros define the sizes and other characteristics of the standard
1506 basic data types used in programs being compiled. Unlike the macros in
1507 the previous section, these apply to specific features of C and related
1508 languages, rather than to fundamental aspects of storage layout.
1510 @defmac INT_TYPE_SIZE
1511 A C expression for the size in bits of the type @code{int} on the
1512 target machine. If you don't define this, the default is one word.
1515 @defmac SHORT_TYPE_SIZE
1516 A C expression for the size in bits of the type @code{short} on the
1517 target machine. If you don't define this, the default is half a word.
1518 (If this would be less than one storage unit, it is rounded up to one
1522 @defmac LONG_TYPE_SIZE
1523 A C expression for the size in bits of the type @code{long} on the
1524 target machine. If you don't define this, the default is one word.
1527 @defmac ADA_LONG_TYPE_SIZE
1528 On some machines, the size used for the Ada equivalent of the type
1529 @code{long} by a native Ada compiler differs from that used by C@. In
1530 that situation, define this macro to be a C expression to be used for
1531 the size of that type. If you don't define this, the default is the
1532 value of @code{LONG_TYPE_SIZE}.
1535 @defmac LONG_LONG_TYPE_SIZE
1536 A C expression for the size in bits of the type @code{long long} on the
1537 target machine. If you don't define this, the default is two
1538 words. If you want to support GNU Ada on your machine, the value of this
1539 macro must be at least 64.
1542 @defmac CHAR_TYPE_SIZE
1543 A C expression for the size in bits of the type @code{char} on the
1544 target machine. If you don't define this, the default is
1545 @code{BITS_PER_UNIT}.
1548 @defmac BOOL_TYPE_SIZE
1549 A C expression for the size in bits of the C++ type @code{bool} and
1550 C99 type @code{_Bool} on the target machine. If you don't define
1551 this, and you probably shouldn't, the default is @code{CHAR_TYPE_SIZE}.
1554 @defmac FLOAT_TYPE_SIZE
1555 A C expression for the size in bits of the type @code{float} on the
1556 target machine. If you don't define this, the default is one word.
1559 @defmac DOUBLE_TYPE_SIZE
1560 A C expression for the size in bits of the type @code{double} on the
1561 target machine. If you don't define this, the default is two
1565 @defmac LONG_DOUBLE_TYPE_SIZE
1566 A C expression for the size in bits of the type @code{long double} on
1567 the target machine. If you don't define this, the default is two
1571 @defmac SHORT_FRACT_TYPE_SIZE
1572 A C expression for the size in bits of the type @code{short _Fract} on
1573 the target machine. If you don't define this, the default is
1574 @code{BITS_PER_UNIT}.
1577 @defmac FRACT_TYPE_SIZE
1578 A C expression for the size in bits of the type @code{_Fract} on
1579 the target machine. If you don't define this, the default is
1580 @code{BITS_PER_UNIT * 2}.
1583 @defmac LONG_FRACT_TYPE_SIZE
1584 A C expression for the size in bits of the type @code{long _Fract} on
1585 the target machine. If you don't define this, the default is
1586 @code{BITS_PER_UNIT * 4}.
1589 @defmac LONG_LONG_FRACT_TYPE_SIZE
1590 A C expression for the size in bits of the type @code{long long _Fract} on
1591 the target machine. If you don't define this, the default is
1592 @code{BITS_PER_UNIT * 8}.
1595 @defmac SHORT_ACCUM_TYPE_SIZE
1596 A C expression for the size in bits of the type @code{short _Accum} on
1597 the target machine. If you don't define this, the default is
1598 @code{BITS_PER_UNIT * 2}.
1601 @defmac ACCUM_TYPE_SIZE
1602 A C expression for the size in bits of the type @code{_Accum} on
1603 the target machine. If you don't define this, the default is
1604 @code{BITS_PER_UNIT * 4}.
1607 @defmac LONG_ACCUM_TYPE_SIZE
1608 A C expression for the size in bits of the type @code{long _Accum} on
1609 the target machine. If you don't define this, the default is
1610 @code{BITS_PER_UNIT * 8}.
1613 @defmac LONG_LONG_ACCUM_TYPE_SIZE
1614 A C expression for the size in bits of the type @code{long long _Accum} on
1615 the target machine. If you don't define this, the default is
1616 @code{BITS_PER_UNIT * 16}.
1619 @defmac LIBGCC2_LONG_DOUBLE_TYPE_SIZE
1620 Define this macro if @code{LONG_DOUBLE_TYPE_SIZE} is not constant or
1621 if you want routines in @file{libgcc2.a} for a size other than
1622 @code{LONG_DOUBLE_TYPE_SIZE}. If you don't define this, the
1623 default is @code{LONG_DOUBLE_TYPE_SIZE}.
1626 @defmac LIBGCC2_HAS_DF_MODE
1627 Define this macro if neither @code{LIBGCC2_DOUBLE_TYPE_SIZE} nor
1628 @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is
1629 @code{DFmode} but you want @code{DFmode} routines in @file{libgcc2.a}
1630 anyway. If you don't define this and either @code{LIBGCC2_DOUBLE_TYPE_SIZE}
1631 or @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is 64 then the default is 1,
1635 @defmac LIBGCC2_HAS_XF_MODE
1636 Define this macro if @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is not
1637 @code{XFmode} but you want @code{XFmode} routines in @file{libgcc2.a}
1638 anyway. If you don't define this and @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE}
1639 is 80 then the default is 1, otherwise it is 0.
1642 @defmac LIBGCC2_HAS_TF_MODE
1643 Define this macro if @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is not
1644 @code{TFmode} but you want @code{TFmode} routines in @file{libgcc2.a}
1645 anyway. If you don't define this and @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE}
1646 is 128 then the default is 1, otherwise it is 0.
1653 Define these macros to be the size in bits of the mantissa of
1654 @code{SFmode}, @code{DFmode}, @code{XFmode} and @code{TFmode} values,
1655 if the defaults in @file{libgcc2.h} are inappropriate. By default,
1656 @code{FLT_MANT_DIG} is used for @code{SF_SIZE}, @code{LDBL_MANT_DIG}
1657 for @code{XF_SIZE} and @code{TF_SIZE}, and @code{DBL_MANT_DIG} or
1658 @code{LDBL_MANT_DIG} for @code{DF_SIZE} according to whether
1659 @code{LIBGCC2_DOUBLE_TYPE_SIZE} or
1660 @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is 64.
1663 @defmac TARGET_FLT_EVAL_METHOD
1664 A C expression for the value for @code{FLT_EVAL_METHOD} in @file{float.h},
1665 assuming, if applicable, that the floating-point control word is in its
1666 default state. If you do not define this macro the value of
1667 @code{FLT_EVAL_METHOD} will be zero.
1670 @defmac WIDEST_HARDWARE_FP_SIZE
1671 A C expression for the size in bits of the widest floating-point format
1672 supported by the hardware. If you define this macro, you must specify a
1673 value less than or equal to the value of @code{LONG_DOUBLE_TYPE_SIZE}.
1674 If you do not define this macro, the value of @code{LONG_DOUBLE_TYPE_SIZE}
1678 @defmac DEFAULT_SIGNED_CHAR
1679 An expression whose value is 1 or 0, according to whether the type
1680 @code{char} should be signed or unsigned by default. The user can
1681 always override this default with the options @option{-fsigned-char}
1682 and @option{-funsigned-char}.
1685 @deftypefn {Target Hook} bool TARGET_DEFAULT_SHORT_ENUMS (void)
1686 This target hook should return true if the compiler should give an
1687 @code{enum} type only as many bytes as it takes to represent the range
1688 of possible values of that type. It should return false if all
1689 @code{enum} types should be allocated like @code{int}.
1691 The default is to return false.
1695 A C expression for a string describing the name of the data type to use
1696 for size values. The typedef name @code{size_t} is defined using the
1697 contents of the string.
1699 The string can contain more than one keyword. If so, separate them with
1700 spaces, and write first any length keyword, then @code{unsigned} if
1701 appropriate, and finally @code{int}. The string must exactly match one
1702 of the data type names defined in the function
1703 @code{init_decl_processing} in the file @file{c-decl.c}. You may not
1704 omit @code{int} or change the order---that would cause the compiler to
1707 If you don't define this macro, the default is @code{"long unsigned
1711 @defmac PTRDIFF_TYPE
1712 A C expression for a string describing the name of the data type to use
1713 for the result of subtracting two pointers. The typedef name
1714 @code{ptrdiff_t} is defined using the contents of the string. See
1715 @code{SIZE_TYPE} above for more information.
1717 If you don't define this macro, the default is @code{"long int"}.
1721 A C expression for a string describing the name of the data type to use
1722 for wide characters. The typedef name @code{wchar_t} is defined using
1723 the contents of the string. See @code{SIZE_TYPE} above for more
1726 If you don't define this macro, the default is @code{"int"}.
1729 @defmac WCHAR_TYPE_SIZE
1730 A C expression for the size in bits of the data type for wide
1731 characters. This is used in @code{cpp}, which cannot make use of
1736 A C expression for a string describing the name of the data type to
1737 use for wide characters passed to @code{printf} and returned from
1738 @code{getwc}. The typedef name @code{wint_t} is defined using the
1739 contents of the string. See @code{SIZE_TYPE} above for more
1742 If you don't define this macro, the default is @code{"unsigned int"}.
1746 A C expression for a string describing the name of the data type that
1747 can represent any value of any standard or extended signed integer type.
1748 The typedef name @code{intmax_t} is defined using the contents of the
1749 string. See @code{SIZE_TYPE} above for more information.
1751 If you don't define this macro, the default is the first of
1752 @code{"int"}, @code{"long int"}, or @code{"long long int"} that has as
1753 much precision as @code{long long int}.
1756 @defmac UINTMAX_TYPE
1757 A C expression for a string describing the name of the data type that
1758 can represent any value of any standard or extended unsigned integer
1759 type. The typedef name @code{uintmax_t} is defined using the contents
1760 of the string. See @code{SIZE_TYPE} above for more information.
1762 If you don't define this macro, the default is the first of
1763 @code{"unsigned int"}, @code{"long unsigned int"}, or @code{"long long
1764 unsigned int"} that has as much precision as @code{long long unsigned
1768 @defmac SIG_ATOMIC_TYPE
1774 @defmacx UINT16_TYPE
1775 @defmacx UINT32_TYPE
1776 @defmacx UINT64_TYPE
1777 @defmacx INT_LEAST8_TYPE
1778 @defmacx INT_LEAST16_TYPE
1779 @defmacx INT_LEAST32_TYPE
1780 @defmacx INT_LEAST64_TYPE
1781 @defmacx UINT_LEAST8_TYPE
1782 @defmacx UINT_LEAST16_TYPE
1783 @defmacx UINT_LEAST32_TYPE
1784 @defmacx UINT_LEAST64_TYPE
1785 @defmacx INT_FAST8_TYPE
1786 @defmacx INT_FAST16_TYPE
1787 @defmacx INT_FAST32_TYPE
1788 @defmacx INT_FAST64_TYPE
1789 @defmacx UINT_FAST8_TYPE
1790 @defmacx UINT_FAST16_TYPE
1791 @defmacx UINT_FAST32_TYPE
1792 @defmacx UINT_FAST64_TYPE
1793 @defmacx INTPTR_TYPE
1794 @defmacx UINTPTR_TYPE
1795 C expressions for the standard types @code{sig_atomic_t},
1796 @code{int8_t}, @code{int16_t}, @code{int32_t}, @code{int64_t},
1797 @code{uint8_t}, @code{uint16_t}, @code{uint32_t}, @code{uint64_t},
1798 @code{int_least8_t}, @code{int_least16_t}, @code{int_least32_t},
1799 @code{int_least64_t}, @code{uint_least8_t}, @code{uint_least16_t},
1800 @code{uint_least32_t}, @code{uint_least64_t}, @code{int_fast8_t},
1801 @code{int_fast16_t}, @code{int_fast32_t}, @code{int_fast64_t},
1802 @code{uint_fast8_t}, @code{uint_fast16_t}, @code{uint_fast32_t},
1803 @code{uint_fast64_t}, @code{intptr_t}, and @code{uintptr_t}. See
1804 @code{SIZE_TYPE} above for more information.
1806 If any of these macros evaluates to a null pointer, the corresponding
1807 type is not supported; if GCC is configured to provide
1808 @code{<stdint.h>} in such a case, the header provided may not conform
1809 to C99, depending on the type in question. The defaults for all of
1810 these macros are null pointers.
1813 @defmac TARGET_PTRMEMFUNC_VBIT_LOCATION
1814 The C++ compiler represents a pointer-to-member-function with a struct
1821 ptrdiff_t vtable_index;
1828 The C++ compiler must use one bit to indicate whether the function that
1829 will be called through a pointer-to-member-function is virtual.
1830 Normally, we assume that the low-order bit of a function pointer must
1831 always be zero. Then, by ensuring that the vtable_index is odd, we can
1832 distinguish which variant of the union is in use. But, on some
1833 platforms function pointers can be odd, and so this doesn't work. In
1834 that case, we use the low-order bit of the @code{delta} field, and shift
1835 the remainder of the @code{delta} field to the left.
1837 GCC will automatically make the right selection about where to store
1838 this bit using the @code{FUNCTION_BOUNDARY} setting for your platform.
1839 However, some platforms such as ARM/Thumb have @code{FUNCTION_BOUNDARY}
1840 set such that functions always start at even addresses, but the lowest
1841 bit of pointers to functions indicate whether the function at that
1842 address is in ARM or Thumb mode. If this is the case of your
1843 architecture, you should define this macro to
1844 @code{ptrmemfunc_vbit_in_delta}.
1846 In general, you should not have to define this macro. On architectures
1847 in which function addresses are always even, according to
1848 @code{FUNCTION_BOUNDARY}, GCC will automatically define this macro to
1849 @code{ptrmemfunc_vbit_in_pfn}.
1852 @defmac TARGET_VTABLE_USES_DESCRIPTORS
1853 Normally, the C++ compiler uses function pointers in vtables. This
1854 macro allows the target to change to use ``function descriptors''
1855 instead. Function descriptors are found on targets for whom a
1856 function pointer is actually a small data structure. Normally the
1857 data structure consists of the actual code address plus a data
1858 pointer to which the function's data is relative.
1860 If vtables are used, the value of this macro should be the number
1861 of words that the function descriptor occupies.
1864 @defmac TARGET_VTABLE_ENTRY_ALIGN
1865 By default, the vtable entries are void pointers, the so the alignment
1866 is the same as pointer alignment. The value of this macro specifies
1867 the alignment of the vtable entry in bits. It should be defined only
1868 when special alignment is necessary. */
1871 @defmac TARGET_VTABLE_DATA_ENTRY_DISTANCE
1872 There are a few non-descriptor entries in the vtable at offsets below
1873 zero. If these entries must be padded (say, to preserve the alignment
1874 specified by @code{TARGET_VTABLE_ENTRY_ALIGN}), set this to the number
1875 of words in each data entry.
1879 @section Register Usage
1880 @cindex register usage
1882 This section explains how to describe what registers the target machine
1883 has, and how (in general) they can be used.
1885 The description of which registers a specific instruction can use is
1886 done with register classes; see @ref{Register Classes}. For information
1887 on using registers to access a stack frame, see @ref{Frame Registers}.
1888 For passing values in registers, see @ref{Register Arguments}.
1889 For returning values in registers, see @ref{Scalar Return}.
1892 * Register Basics:: Number and kinds of registers.
1893 * Allocation Order:: Order in which registers are allocated.
1894 * Values in Registers:: What kinds of values each reg can hold.
1895 * Leaf Functions:: Renumbering registers for leaf functions.
1896 * Stack Registers:: Handling a register stack such as 80387.
1899 @node Register Basics
1900 @subsection Basic Characteristics of Registers
1902 @c prevent bad page break with this line
1903 Registers have various characteristics.
1905 @defmac FIRST_PSEUDO_REGISTER
1906 Number of hardware registers known to the compiler. They receive
1907 numbers 0 through @code{FIRST_PSEUDO_REGISTER-1}; thus, the first
1908 pseudo register's number really is assigned the number
1909 @code{FIRST_PSEUDO_REGISTER}.
1912 @defmac FIXED_REGISTERS
1913 @cindex fixed register
1914 An initializer that says which registers are used for fixed purposes
1915 all throughout the compiled code and are therefore not available for
1916 general allocation. These would include the stack pointer, the frame
1917 pointer (except on machines where that can be used as a general
1918 register when no frame pointer is needed), the program counter on
1919 machines where that is considered one of the addressable registers,
1920 and any other numbered register with a standard use.
1922 This information is expressed as a sequence of numbers, separated by
1923 commas and surrounded by braces. The @var{n}th number is 1 if
1924 register @var{n} is fixed, 0 otherwise.
1926 The table initialized from this macro, and the table initialized by
1927 the following one, may be overridden at run time either automatically,
1928 by the actions of the macro @code{CONDITIONAL_REGISTER_USAGE}, or by
1929 the user with the command options @option{-ffixed-@var{reg}},
1930 @option{-fcall-used-@var{reg}} and @option{-fcall-saved-@var{reg}}.
1933 @defmac CALL_USED_REGISTERS
1934 @cindex call-used register
1935 @cindex call-clobbered register
1936 @cindex call-saved register
1937 Like @code{FIXED_REGISTERS} but has 1 for each register that is
1938 clobbered (in general) by function calls as well as for fixed
1939 registers. This macro therefore identifies the registers that are not
1940 available for general allocation of values that must live across
1943 If a register has 0 in @code{CALL_USED_REGISTERS}, the compiler
1944 automatically saves it on function entry and restores it on function
1945 exit, if the register is used within the function.
1948 @defmac CALL_REALLY_USED_REGISTERS
1949 @cindex call-used register
1950 @cindex call-clobbered register
1951 @cindex call-saved register
1952 Like @code{CALL_USED_REGISTERS} except this macro doesn't require
1953 that the entire set of @code{FIXED_REGISTERS} be included.
1954 (@code{CALL_USED_REGISTERS} must be a superset of @code{FIXED_REGISTERS}).
1955 This macro is optional. If not specified, it defaults to the value
1956 of @code{CALL_USED_REGISTERS}.
1959 @defmac HARD_REGNO_CALL_PART_CLOBBERED (@var{regno}, @var{mode})
1960 @cindex call-used register
1961 @cindex call-clobbered register
1962 @cindex call-saved register
1963 A C expression that is nonzero if it is not permissible to store a
1964 value of mode @var{mode} in hard register number @var{regno} across a
1965 call without some part of it being clobbered. For most machines this
1966 macro need not be defined. It is only required for machines that do not
1967 preserve the entire contents of a register across a call.
1971 @findex call_used_regs
1974 @findex reg_class_contents
1975 @defmac CONDITIONAL_REGISTER_USAGE
1976 Zero or more C statements that may conditionally modify five variables
1977 @code{fixed_regs}, @code{call_used_regs}, @code{global_regs},
1978 @code{reg_names}, and @code{reg_class_contents}, to take into account
1979 any dependence of these register sets on target flags. The first three
1980 of these are of type @code{char []} (interpreted as Boolean vectors).
1981 @code{global_regs} is a @code{const char *[]}, and
1982 @code{reg_class_contents} is a @code{HARD_REG_SET}. Before the macro is
1983 called, @code{fixed_regs}, @code{call_used_regs},
1984 @code{reg_class_contents}, and @code{reg_names} have been initialized
1985 from @code{FIXED_REGISTERS}, @code{CALL_USED_REGISTERS},
1986 @code{REG_CLASS_CONTENTS}, and @code{REGISTER_NAMES}, respectively.
1987 @code{global_regs} has been cleared, and any @option{-ffixed-@var{reg}},
1988 @option{-fcall-used-@var{reg}} and @option{-fcall-saved-@var{reg}}
1989 command options have been applied.
1991 You need not define this macro if it has no work to do.
1993 @cindex disabling certain registers
1994 @cindex controlling register usage
1995 If the usage of an entire class of registers depends on the target
1996 flags, you may indicate this to GCC by using this macro to modify
1997 @code{fixed_regs} and @code{call_used_regs} to 1 for each of the
1998 registers in the classes which should not be used by GCC@. Also define
1999 the macro @code{REG_CLASS_FROM_LETTER} / @code{REG_CLASS_FROM_CONSTRAINT}
2000 to return @code{NO_REGS} if it
2001 is called with a letter for a class that shouldn't be used.
2003 (However, if this class is not included in @code{GENERAL_REGS} and all
2004 of the insn patterns whose constraints permit this class are
2005 controlled by target switches, then GCC will automatically avoid using
2006 these registers when the target switches are opposed to them.)
2009 @defmac INCOMING_REGNO (@var{out})
2010 Define this macro if the target machine has register windows. This C
2011 expression returns the register number as seen by the called function
2012 corresponding to the register number @var{out} as seen by the calling
2013 function. Return @var{out} if register number @var{out} is not an
2017 @defmac OUTGOING_REGNO (@var{in})
2018 Define this macro if the target machine has register windows. This C
2019 expression returns the register number as seen by the calling function
2020 corresponding to the register number @var{in} as seen by the called
2021 function. Return @var{in} if register number @var{in} is not an inbound
2025 @defmac LOCAL_REGNO (@var{regno})
2026 Define this macro if the target machine has register windows. This C
2027 expression returns true if the register is call-saved but is in the
2028 register window. Unlike most call-saved registers, such registers
2029 need not be explicitly restored on function exit or during non-local
2034 If the program counter has a register number, define this as that
2035 register number. Otherwise, do not define it.
2038 @node Allocation Order
2039 @subsection Order of Allocation of Registers
2040 @cindex order of register allocation
2041 @cindex register allocation order
2043 @c prevent bad page break with this line
2044 Registers are allocated in order.
2046 @defmac REG_ALLOC_ORDER
2047 If defined, an initializer for a vector of integers, containing the
2048 numbers of hard registers in the order in which GCC should prefer
2049 to use them (from most preferred to least).
2051 If this macro is not defined, registers are used lowest numbered first
2052 (all else being equal).
2054 One use of this macro is on machines where the highest numbered
2055 registers must always be saved and the save-multiple-registers
2056 instruction supports only sequences of consecutive registers. On such
2057 machines, define @code{REG_ALLOC_ORDER} to be an initializer that lists
2058 the highest numbered allocable register first.
2061 @defmac ADJUST_REG_ALLOC_ORDER
2062 A C statement (sans semicolon) to choose the order in which to allocate
2063 hard registers for pseudo-registers local to a basic block.
2065 Store the desired register order in the array @code{reg_alloc_order}.
2066 Element 0 should be the register to allocate first; element 1, the next
2067 register; and so on.
2069 The macro body should not assume anything about the contents of
2070 @code{reg_alloc_order} before execution of the macro.
2072 On most machines, it is not necessary to define this macro.
2075 @defmac HONOR_REG_ALLOC_ORDER
2076 Normally, IRA tries to estimate the costs for saving a register in the
2077 prologue and restoring it in the epilogue. This discourages it from
2078 using call-saved registers. If a machine wants to ensure that IRA
2079 allocates registers in the order given by REG_ALLOC_ORDER even if some
2080 call-saved registers appear earlier than call-used ones, this macro
2084 @defmac IRA_HARD_REGNO_ADD_COST_MULTIPLIER (@var{regno})
2085 In some case register allocation order is not enough for the
2086 Integrated Register Allocator (@acronym{IRA}) to generate a good code.
2087 If this macro is defined, it should return a floating point value
2088 based on @var{regno}. The cost of using @var{regno} for a pseudo will
2089 be increased by approximately the pseudo's usage frequency times the
2090 value returned by this macro. Not defining this macro is equivalent
2091 to having it always return @code{0.0}.
2093 On most machines, it is not necessary to define this macro.
2096 @node Values in Registers
2097 @subsection How Values Fit in Registers
2099 This section discusses the macros that describe which kinds of values
2100 (specifically, which machine modes) each register can hold, and how many
2101 consecutive registers are needed for a given mode.
2103 @defmac HARD_REGNO_NREGS (@var{regno}, @var{mode})
2104 A C expression for the number of consecutive hard registers, starting
2105 at register number @var{regno}, required to hold a value of mode
2106 @var{mode}. This macro must never return zero, even if a register
2107 cannot hold the requested mode - indicate that with HARD_REGNO_MODE_OK
2108 and/or CANNOT_CHANGE_MODE_CLASS instead.
2110 On a machine where all registers are exactly one word, a suitable
2111 definition of this macro is
2114 #define HARD_REGNO_NREGS(REGNO, MODE) \
2115 ((GET_MODE_SIZE (MODE) + UNITS_PER_WORD - 1) \
2120 @defmac HARD_REGNO_NREGS_HAS_PADDING (@var{regno}, @var{mode})
2121 A C expression that is nonzero if a value of mode @var{mode}, stored
2122 in memory, ends with padding that causes it to take up more space than
2123 in registers starting at register number @var{regno} (as determined by
2124 multiplying GCC's notion of the size of the register when containing
2125 this mode by the number of registers returned by
2126 @code{HARD_REGNO_NREGS}). By default this is zero.
2128 For example, if a floating-point value is stored in three 32-bit
2129 registers but takes up 128 bits in memory, then this would be
2132 This macros only needs to be defined if there are cases where
2133 @code{subreg_get_info}
2134 would otherwise wrongly determine that a @code{subreg} can be
2135 represented by an offset to the register number, when in fact such a
2136 @code{subreg} would contain some of the padding not stored in
2137 registers and so not be representable.
2140 @defmac HARD_REGNO_NREGS_WITH_PADDING (@var{regno}, @var{mode})
2141 For values of @var{regno} and @var{mode} for which
2142 @code{HARD_REGNO_NREGS_HAS_PADDING} returns nonzero, a C expression
2143 returning the greater number of registers required to hold the value
2144 including any padding. In the example above, the value would be four.
2147 @defmac REGMODE_NATURAL_SIZE (@var{mode})
2148 Define this macro if the natural size of registers that hold values
2149 of mode @var{mode} is not the word size. It is a C expression that
2150 should give the natural size in bytes for the specified mode. It is
2151 used by the register allocator to try to optimize its results. This
2152 happens for example on SPARC 64-bit where the natural size of
2153 floating-point registers is still 32-bit.
2156 @defmac HARD_REGNO_MODE_OK (@var{regno}, @var{mode})
2157 A C expression that is nonzero if it is permissible to store a value
2158 of mode @var{mode} in hard register number @var{regno} (or in several
2159 registers starting with that one). For a machine where all registers
2160 are equivalent, a suitable definition is
2163 #define HARD_REGNO_MODE_OK(REGNO, MODE) 1
2166 You need not include code to check for the numbers of fixed registers,
2167 because the allocation mechanism considers them to be always occupied.
2169 @cindex register pairs
2170 On some machines, double-precision values must be kept in even/odd
2171 register pairs. You can implement that by defining this macro to reject
2172 odd register numbers for such modes.
2174 The minimum requirement for a mode to be OK in a register is that the
2175 @samp{mov@var{mode}} instruction pattern support moves between the
2176 register and other hard register in the same class and that moving a
2177 value into the register and back out not alter it.
2179 Since the same instruction used to move @code{word_mode} will work for
2180 all narrower integer modes, it is not necessary on any machine for
2181 @code{HARD_REGNO_MODE_OK} to distinguish between these modes, provided
2182 you define patterns @samp{movhi}, etc., to take advantage of this. This
2183 is useful because of the interaction between @code{HARD_REGNO_MODE_OK}
2184 and @code{MODES_TIEABLE_P}; it is very desirable for all integer modes
2187 Many machines have special registers for floating point arithmetic.
2188 Often people assume that floating point machine modes are allowed only
2189 in floating point registers. This is not true. Any registers that
2190 can hold integers can safely @emph{hold} a floating point machine
2191 mode, whether or not floating arithmetic can be done on it in those
2192 registers. Integer move instructions can be used to move the values.
2194 On some machines, though, the converse is true: fixed-point machine
2195 modes may not go in floating registers. This is true if the floating
2196 registers normalize any value stored in them, because storing a
2197 non-floating value there would garble it. In this case,
2198 @code{HARD_REGNO_MODE_OK} should reject fixed-point machine modes in
2199 floating registers. But if the floating registers do not automatically
2200 normalize, if you can store any bit pattern in one and retrieve it
2201 unchanged without a trap, then any machine mode may go in a floating
2202 register, so you can define this macro to say so.
2204 The primary significance of special floating registers is rather that
2205 they are the registers acceptable in floating point arithmetic
2206 instructions. However, this is of no concern to
2207 @code{HARD_REGNO_MODE_OK}. You handle it by writing the proper
2208 constraints for those instructions.
2210 On some machines, the floating registers are especially slow to access,
2211 so that it is better to store a value in a stack frame than in such a
2212 register if floating point arithmetic is not being done. As long as the
2213 floating registers are not in class @code{GENERAL_REGS}, they will not
2214 be used unless some pattern's constraint asks for one.
2217 @defmac HARD_REGNO_RENAME_OK (@var{from}, @var{to})
2218 A C expression that is nonzero if it is OK to rename a hard register
2219 @var{from} to another hard register @var{to}.
2221 One common use of this macro is to prevent renaming of a register to
2222 another register that is not saved by a prologue in an interrupt
2225 The default is always nonzero.
2228 @defmac MODES_TIEABLE_P (@var{mode1}, @var{mode2})
2229 A C expression that is nonzero if a value of mode
2230 @var{mode1} is accessible in mode @var{mode2} without copying.
2232 If @code{HARD_REGNO_MODE_OK (@var{r}, @var{mode1})} and
2233 @code{HARD_REGNO_MODE_OK (@var{r}, @var{mode2})} are always the same for
2234 any @var{r}, then @code{MODES_TIEABLE_P (@var{mode1}, @var{mode2})}
2235 should be nonzero. If they differ for any @var{r}, you should define
2236 this macro to return zero unless some other mechanism ensures the
2237 accessibility of the value in a narrower mode.
2239 You should define this macro to return nonzero in as many cases as
2240 possible since doing so will allow GCC to perform better register
2244 @deftypefn {Target Hook} bool TARGET_HARD_REGNO_SCRATCH_OK (unsigned int @var{regno})
2245 This target hook should return @code{true} if it is OK to use a hard register
2246 @var{regno} as scratch reg in peephole2.
2248 One common use of this macro is to prevent using of a register that
2249 is not saved by a prologue in an interrupt handler.
2251 The default version of this hook always returns @code{true}.
2254 @defmac AVOID_CCMODE_COPIES
2255 Define this macro if the compiler should avoid copies to/from @code{CCmode}
2256 registers. You should only define this macro if support for copying to/from
2257 @code{CCmode} is incomplete.
2260 @node Leaf Functions
2261 @subsection Handling Leaf Functions
2263 @cindex leaf functions
2264 @cindex functions, leaf
2265 On some machines, a leaf function (i.e., one which makes no calls) can run
2266 more efficiently if it does not make its own register window. Often this
2267 means it is required to receive its arguments in the registers where they
2268 are passed by the caller, instead of the registers where they would
2271 The special treatment for leaf functions generally applies only when
2272 other conditions are met; for example, often they may use only those
2273 registers for its own variables and temporaries. We use the term ``leaf
2274 function'' to mean a function that is suitable for this special
2275 handling, so that functions with no calls are not necessarily ``leaf
2278 GCC assigns register numbers before it knows whether the function is
2279 suitable for leaf function treatment. So it needs to renumber the
2280 registers in order to output a leaf function. The following macros
2283 @defmac LEAF_REGISTERS
2284 Name of a char vector, indexed by hard register number, which
2285 contains 1 for a register that is allowable in a candidate for leaf
2288 If leaf function treatment involves renumbering the registers, then the
2289 registers marked here should be the ones before renumbering---those that
2290 GCC would ordinarily allocate. The registers which will actually be
2291 used in the assembler code, after renumbering, should not be marked with 1
2294 Define this macro only if the target machine offers a way to optimize
2295 the treatment of leaf functions.
2298 @defmac LEAF_REG_REMAP (@var{regno})
2299 A C expression whose value is the register number to which @var{regno}
2300 should be renumbered, when a function is treated as a leaf function.
2302 If @var{regno} is a register number which should not appear in a leaf
2303 function before renumbering, then the expression should yield @minus{}1, which
2304 will cause the compiler to abort.
2306 Define this macro only if the target machine offers a way to optimize the
2307 treatment of leaf functions, and registers need to be renumbered to do
2311 @findex current_function_is_leaf
2312 @findex current_function_uses_only_leaf_regs
2313 @code{TARGET_ASM_FUNCTION_PROLOGUE} and
2314 @code{TARGET_ASM_FUNCTION_EPILOGUE} must usually treat leaf functions
2315 specially. They can test the C variable @code{current_function_is_leaf}
2316 which is nonzero for leaf functions. @code{current_function_is_leaf} is
2317 set prior to local register allocation and is valid for the remaining
2318 compiler passes. They can also test the C variable
2319 @code{current_function_uses_only_leaf_regs} which is nonzero for leaf
2320 functions which only use leaf registers.
2321 @code{current_function_uses_only_leaf_regs} is valid after all passes
2322 that modify the instructions have been run and is only useful if
2323 @code{LEAF_REGISTERS} is defined.
2324 @c changed this to fix overfull. ALSO: why the "it" at the beginning
2325 @c of the next paragraph?! --mew 2feb93
2327 @node Stack Registers
2328 @subsection Registers That Form a Stack
2330 There are special features to handle computers where some of the
2331 ``registers'' form a stack. Stack registers are normally written by
2332 pushing onto the stack, and are numbered relative to the top of the
2335 Currently, GCC can only handle one group of stack-like registers, and
2336 they must be consecutively numbered. Furthermore, the existing
2337 support for stack-like registers is specific to the 80387 floating
2338 point coprocessor. If you have a new architecture that uses
2339 stack-like registers, you will need to do substantial work on
2340 @file{reg-stack.c} and write your machine description to cooperate
2341 with it, as well as defining these macros.
2344 Define this if the machine has any stack-like registers.
2347 @defmac STACK_REG_COVER_CLASS
2348 This is a cover class containing the stack registers. Define this if
2349 the machine has any stack-like registers.
2352 @defmac FIRST_STACK_REG
2353 The number of the first stack-like register. This one is the top
2357 @defmac LAST_STACK_REG
2358 The number of the last stack-like register. This one is the bottom of
2362 @node Register Classes
2363 @section Register Classes
2364 @cindex register class definitions
2365 @cindex class definitions, register
2367 On many machines, the numbered registers are not all equivalent.
2368 For example, certain registers may not be allowed for indexed addressing;
2369 certain registers may not be allowed in some instructions. These machine
2370 restrictions are described to the compiler using @dfn{register classes}.
2372 You define a number of register classes, giving each one a name and saying
2373 which of the registers belong to it. Then you can specify register classes
2374 that are allowed as operands to particular instruction patterns.
2378 In general, each register will belong to several classes. In fact, one
2379 class must be named @code{ALL_REGS} and contain all the registers. Another
2380 class must be named @code{NO_REGS} and contain no registers. Often the
2381 union of two classes will be another class; however, this is not required.
2383 @findex GENERAL_REGS
2384 One of the classes must be named @code{GENERAL_REGS}. There is nothing
2385 terribly special about the name, but the operand constraint letters
2386 @samp{r} and @samp{g} specify this class. If @code{GENERAL_REGS} is
2387 the same as @code{ALL_REGS}, just define it as a macro which expands
2390 Order the classes so that if class @var{x} is contained in class @var{y}
2391 then @var{x} has a lower class number than @var{y}.
2393 The way classes other than @code{GENERAL_REGS} are specified in operand
2394 constraints is through machine-dependent operand constraint letters.
2395 You can define such letters to correspond to various classes, then use
2396 them in operand constraints.
2398 You should define a class for the union of two classes whenever some
2399 instruction allows both classes. For example, if an instruction allows
2400 either a floating point (coprocessor) register or a general register for a
2401 certain operand, you should define a class @code{FLOAT_OR_GENERAL_REGS}
2402 which includes both of them. Otherwise you will get suboptimal code.
2404 You must also specify certain redundant information about the register
2405 classes: for each class, which classes contain it and which ones are
2406 contained in it; for each pair of classes, the largest class contained
2409 When a value occupying several consecutive registers is expected in a
2410 certain class, all the registers used must belong to that class.
2411 Therefore, register classes cannot be used to enforce a requirement for
2412 a register pair to start with an even-numbered register. The way to
2413 specify this requirement is with @code{HARD_REGNO_MODE_OK}.
2415 Register classes used for input-operands of bitwise-and or shift
2416 instructions have a special requirement: each such class must have, for
2417 each fixed-point machine mode, a subclass whose registers can transfer that
2418 mode to or from memory. For example, on some machines, the operations for
2419 single-byte values (@code{QImode}) are limited to certain registers. When
2420 this is so, each register class that is used in a bitwise-and or shift
2421 instruction must have a subclass consisting of registers from which
2422 single-byte values can be loaded or stored. This is so that
2423 @code{PREFERRED_RELOAD_CLASS} can always have a possible value to return.
2425 @deftp {Data type} {enum reg_class}
2426 An enumerated type that must be defined with all the register class names
2427 as enumerated values. @code{NO_REGS} must be first. @code{ALL_REGS}
2428 must be the last register class, followed by one more enumerated value,
2429 @code{LIM_REG_CLASSES}, which is not a register class but rather
2430 tells how many classes there are.
2432 Each register class has a number, which is the value of casting
2433 the class name to type @code{int}. The number serves as an index
2434 in many of the tables described below.
2437 @defmac N_REG_CLASSES
2438 The number of distinct register classes, defined as follows:
2441 #define N_REG_CLASSES (int) LIM_REG_CLASSES
2445 @defmac REG_CLASS_NAMES
2446 An initializer containing the names of the register classes as C string
2447 constants. These names are used in writing some of the debugging dumps.
2450 @defmac REG_CLASS_CONTENTS
2451 An initializer containing the contents of the register classes, as integers
2452 which are bit masks. The @var{n}th integer specifies the contents of class
2453 @var{n}. The way the integer @var{mask} is interpreted is that
2454 register @var{r} is in the class if @code{@var{mask} & (1 << @var{r})} is 1.
2456 When the machine has more than 32 registers, an integer does not suffice.
2457 Then the integers are replaced by sub-initializers, braced groupings containing
2458 several integers. Each sub-initializer must be suitable as an initializer
2459 for the type @code{HARD_REG_SET} which is defined in @file{hard-reg-set.h}.
2460 In this situation, the first integer in each sub-initializer corresponds to
2461 registers 0 through 31, the second integer to registers 32 through 63, and
2465 @defmac REGNO_REG_CLASS (@var{regno})
2466 A C expression whose value is a register class containing hard register
2467 @var{regno}. In general there is more than one such class; choose a class
2468 which is @dfn{minimal}, meaning that no smaller class also contains the
2472 @defmac BASE_REG_CLASS
2473 A macro whose definition is the name of the class to which a valid
2474 base register must belong. A base register is one used in an address
2475 which is the register value plus a displacement.
2478 @defmac MODE_BASE_REG_CLASS (@var{mode})
2479 This is a variation of the @code{BASE_REG_CLASS} macro which allows
2480 the selection of a base register in a mode dependent manner. If
2481 @var{mode} is VOIDmode then it should return the same value as
2482 @code{BASE_REG_CLASS}.
2485 @defmac MODE_BASE_REG_REG_CLASS (@var{mode})
2486 A C expression whose value is the register class to which a valid
2487 base register must belong in order to be used in a base plus index
2488 register address. You should define this macro if base plus index
2489 addresses have different requirements than other base register uses.
2492 @defmac MODE_CODE_BASE_REG_CLASS (@var{mode}, @var{outer_code}, @var{index_code})
2493 A C expression whose value is the register class to which a valid
2494 base register must belong. @var{outer_code} and @var{index_code} define the
2495 context in which the base register occurs. @var{outer_code} is the code of
2496 the immediately enclosing expression (@code{MEM} for the top level of an
2497 address, @code{ADDRESS} for something that occurs in an
2498 @code{address_operand}). @var{index_code} is the code of the corresponding
2499 index expression if @var{outer_code} is @code{PLUS}; @code{SCRATCH} otherwise.
2502 @defmac INDEX_REG_CLASS
2503 A macro whose definition is the name of the class to which a valid
2504 index register must belong. An index register is one used in an
2505 address where its value is either multiplied by a scale factor or
2506 added to another register (as well as added to a displacement).
2509 @defmac REGNO_OK_FOR_BASE_P (@var{num})
2510 A C expression which is nonzero if register number @var{num} is
2511 suitable for use as a base register in operand addresses.
2512 Like @code{TARGET_LEGITIMATE_ADDRESS_P}, this macro should also
2513 define a strict and a non-strict variant. Both variants behave
2514 the same for hard register; for pseudos, the strict variant will
2515 pass only those that have been allocated to a valid hard registers,
2516 while the non-strict variant will pass all pseudos.
2518 @findex REG_OK_STRICT
2519 Compiler source files that want to use the strict variant of this and
2520 other macros define the macro @code{REG_OK_STRICT}. You should use an
2521 @code{#ifdef REG_OK_STRICT} conditional to define the strict variant in
2522 that case and the non-strict variant otherwise.
2525 @defmac REGNO_MODE_OK_FOR_BASE_P (@var{num}, @var{mode})
2526 A C expression that is just like @code{REGNO_OK_FOR_BASE_P}, except that
2527 that expression may examine the mode of the memory reference in
2528 @var{mode}. You should define this macro if the mode of the memory
2529 reference affects whether a register may be used as a base register. If
2530 you define this macro, the compiler will use it instead of
2531 @code{REGNO_OK_FOR_BASE_P}. The mode may be @code{VOIDmode} for
2532 addresses that appear outside a @code{MEM}, i.e., as an
2533 @code{address_operand}.
2535 This macro also has strict and non-strict variants.
2538 @defmac REGNO_MODE_OK_FOR_REG_BASE_P (@var{num}, @var{mode})
2539 A C expression which is nonzero if register number @var{num} is suitable for
2540 use as a base register in base plus index operand addresses, accessing
2541 memory in mode @var{mode}. It may be either a suitable hard register or a
2542 pseudo register that has been allocated such a hard register. You should
2543 define this macro if base plus index addresses have different requirements
2544 than other base register uses.
2546 Use of this macro is deprecated; please use the more general
2547 @code{REGNO_MODE_CODE_OK_FOR_BASE_P}.
2549 This macro also has strict and non-strict variants.
2552 @defmac REGNO_MODE_CODE_OK_FOR_BASE_P (@var{num}, @var{mode}, @var{outer_code}, @var{index_code})
2553 A C expression that is just like @code{REGNO_MODE_OK_FOR_BASE_P}, except
2554 that that expression may examine the context in which the register
2555 appears in the memory reference. @var{outer_code} is the code of the
2556 immediately enclosing expression (@code{MEM} if at the top level of the
2557 address, @code{ADDRESS} for something that occurs in an
2558 @code{address_operand}). @var{index_code} is the code of the
2559 corresponding index expression if @var{outer_code} is @code{PLUS};
2560 @code{SCRATCH} otherwise. The mode may be @code{VOIDmode} for addresses
2561 that appear outside a @code{MEM}, i.e., as an @code{address_operand}.
2563 This macro also has strict and non-strict variants.
2566 @defmac REGNO_OK_FOR_INDEX_P (@var{num})
2567 A C expression which is nonzero if register number @var{num} is
2568 suitable for use as an index register in operand addresses. It may be
2569 either a suitable hard register or a pseudo register that has been
2570 allocated such a hard register.
2572 The difference between an index register and a base register is that
2573 the index register may be scaled. If an address involves the sum of
2574 two registers, neither one of them scaled, then either one may be
2575 labeled the ``base'' and the other the ``index''; but whichever
2576 labeling is used must fit the machine's constraints of which registers
2577 may serve in each capacity. The compiler will try both labelings,
2578 looking for one that is valid, and will reload one or both registers
2579 only if neither labeling works.
2581 This macro also has strict and non-strict variants.
2584 @defmac PREFERRED_RELOAD_CLASS (@var{x}, @var{class})
2585 A C expression that places additional restrictions on the register class
2586 to use when it is necessary to copy value @var{x} into a register in class
2587 @var{class}. The value is a register class; perhaps @var{class}, or perhaps
2588 another, smaller class. On many machines, the following definition is
2592 #define PREFERRED_RELOAD_CLASS(X,CLASS) CLASS
2595 Sometimes returning a more restrictive class makes better code. For
2596 example, on the 68000, when @var{x} is an integer constant that is in range
2597 for a @samp{moveq} instruction, the value of this macro is always
2598 @code{DATA_REGS} as long as @var{class} includes the data registers.
2599 Requiring a data register guarantees that a @samp{moveq} will be used.
2601 One case where @code{PREFERRED_RELOAD_CLASS} must not return
2602 @var{class} is if @var{x} is a legitimate constant which cannot be
2603 loaded into some register class. By returning @code{NO_REGS} you can
2604 force @var{x} into a memory location. For example, rs6000 can load
2605 immediate values into general-purpose registers, but does not have an
2606 instruction for loading an immediate value into a floating-point
2607 register, so @code{PREFERRED_RELOAD_CLASS} returns @code{NO_REGS} when
2608 @var{x} is a floating-point constant. If the constant can't be loaded
2609 into any kind of register, code generation will be better if
2610 @code{LEGITIMATE_CONSTANT_P} makes the constant illegitimate instead
2611 of using @code{PREFERRED_RELOAD_CLASS}.
2613 If an insn has pseudos in it after register allocation, reload will go
2614 through the alternatives and call repeatedly @code{PREFERRED_RELOAD_CLASS}
2615 to find the best one. Returning @code{NO_REGS}, in this case, makes
2616 reload add a @code{!} in front of the constraint: the x86 back-end uses
2617 this feature to discourage usage of 387 registers when math is done in
2618 the SSE registers (and vice versa).
2621 @defmac PREFERRED_OUTPUT_RELOAD_CLASS (@var{x}, @var{class})
2622 Like @code{PREFERRED_RELOAD_CLASS}, but for output reloads instead of
2623 input reloads. If you don't define this macro, the default is to use
2624 @var{class}, unchanged.
2626 You can also use @code{PREFERRED_OUTPUT_RELOAD_CLASS} to discourage
2627 reload from using some alternatives, like @code{PREFERRED_RELOAD_CLASS}.
2630 @defmac LIMIT_RELOAD_CLASS (@var{mode}, @var{class})
2631 A C expression that places additional restrictions on the register class
2632 to use when it is necessary to be able to hold a value of mode
2633 @var{mode} in a reload register for which class @var{class} would
2636 Unlike @code{PREFERRED_RELOAD_CLASS}, this macro should be used when
2637 there are certain modes that simply can't go in certain reload classes.
2639 The value is a register class; perhaps @var{class}, or perhaps another,
2642 Don't define this macro unless the target machine has limitations which
2643 require the macro to do something nontrivial.
2646 @deftypefn {Target Hook} reg_class_t TARGET_SECONDARY_RELOAD (bool @var{in_p}, rtx @var{x}, reg_class_t @var{reload_class}, enum machine_mode @var{reload_mode}, secondary_reload_info *@var{sri})
2647 Many machines have some registers that cannot be copied directly to or
2648 from memory or even from other types of registers. An example is the
2649 @samp{MQ} register, which on most machines, can only be copied to or
2650 from general registers, but not memory. Below, we shall be using the
2651 term 'intermediate register' when a move operation cannot be performed
2652 directly, but has to be done by copying the source into the intermediate
2653 register first, and then copying the intermediate register to the
2654 destination. An intermediate register always has the same mode as
2655 source and destination. Since it holds the actual value being copied,
2656 reload might apply optimizations to re-use an intermediate register
2657 and eliding the copy from the source when it can determine that the
2658 intermediate register still holds the required value.
2660 Another kind of secondary reload is required on some machines which
2661 allow copying all registers to and from memory, but require a scratch
2662 register for stores to some memory locations (e.g., those with symbolic
2663 address on the RT, and those with certain symbolic address on the SPARC
2664 when compiling PIC)@. Scratch registers need not have the same mode
2665 as the value being copied, and usually hold a different value than
2666 that being copied. Special patterns in the md file are needed to
2667 describe how the copy is performed with the help of the scratch register;
2668 these patterns also describe the number, register class(es) and mode(s)
2669 of the scratch register(s).
2671 In some cases, both an intermediate and a scratch register are required.
2673 For input reloads, this target hook is called with nonzero @var{in_p},
2674 and @var{x} is an rtx that needs to be copied to a register of class
2675 @var{reload_class} in @var{reload_mode}. For output reloads, this target
2676 hook is called with zero @var{in_p}, and a register of class @var{reload_class}
2677 needs to be copied to rtx @var{x} in @var{reload_mode}.
2679 If copying a register of @var{reload_class} from/to @var{x} requires
2680 an intermediate register, the hook @code{secondary_reload} should
2681 return the register class required for this intermediate register.
2682 If no intermediate register is required, it should return NO_REGS.
2683 If more than one intermediate register is required, describe the one
2684 that is closest in the copy chain to the reload register.
2686 If scratch registers are needed, you also have to describe how to
2687 perform the copy from/to the reload register to/from this
2688 closest intermediate register. Or if no intermediate register is
2689 required, but still a scratch register is needed, describe the
2690 copy from/to the reload register to/from the reload operand @var{x}.
2692 You do this by setting @code{sri->icode} to the instruction code of a pattern
2693 in the md file which performs the move. Operands 0 and 1 are the output
2694 and input of this copy, respectively. Operands from operand 2 onward are
2695 for scratch operands. These scratch operands must have a mode, and a
2696 single-register-class
2697 @c [later: or memory]
2700 When an intermediate register is used, the @code{secondary_reload}
2701 hook will be called again to determine how to copy the intermediate
2702 register to/from the reload operand @var{x}, so your hook must also
2703 have code to handle the register class of the intermediate operand.
2705 @c [For later: maybe we'll allow multi-alternative reload patterns -
2706 @c the port maintainer could name a mov<mode> pattern that has clobbers -
2707 @c and match the constraints of input and output to determine the required
2708 @c alternative. A restriction would be that constraints used to match
2709 @c against reloads registers would have to be written as register class
2710 @c constraints, or we need a new target macro / hook that tells us if an
2711 @c arbitrary constraint can match an unknown register of a given class.
2712 @c Such a macro / hook would also be useful in other places.]
2715 @var{x} might be a pseudo-register or a @code{subreg} of a
2716 pseudo-register, which could either be in a hard register or in memory.
2717 Use @code{true_regnum} to find out; it will return @minus{}1 if the pseudo is
2718 in memory and the hard register number if it is in a register.
2720 Scratch operands in memory (constraint @code{"=m"} / @code{"=&m"}) are
2721 currently not supported. For the time being, you will have to continue
2722 to use @code{SECONDARY_MEMORY_NEEDED} for that purpose.
2724 @code{copy_cost} also uses this target hook to find out how values are
2725 copied. If you want it to include some extra cost for the need to allocate
2726 (a) scratch register(s), set @code{sri->extra_cost} to the additional cost.
2727 Or if two dependent moves are supposed to have a lower cost than the sum
2728 of the individual moves due to expected fortuitous scheduling and/or special
2729 forwarding logic, you can set @code{sri->extra_cost} to a negative amount.
2732 @defmac SECONDARY_RELOAD_CLASS (@var{class}, @var{mode}, @var{x})
2733 @defmacx SECONDARY_INPUT_RELOAD_CLASS (@var{class}, @var{mode}, @var{x})
2734 @defmacx SECONDARY_OUTPUT_RELOAD_CLASS (@var{class}, @var{mode}, @var{x})
2735 These macros are obsolete, new ports should use the target hook
2736 @code{TARGET_SECONDARY_RELOAD} instead.
2738 These are obsolete macros, replaced by the @code{TARGET_SECONDARY_RELOAD}
2739 target hook. Older ports still define these macros to indicate to the
2740 reload phase that it may
2741 need to allocate at least one register for a reload in addition to the
2742 register to contain the data. Specifically, if copying @var{x} to a
2743 register @var{class} in @var{mode} requires an intermediate register,
2744 you were supposed to define @code{SECONDARY_INPUT_RELOAD_CLASS} to return the
2745 largest register class all of whose registers can be used as
2746 intermediate registers or scratch registers.
2748 If copying a register @var{class} in @var{mode} to @var{x} requires an
2749 intermediate or scratch register, @code{SECONDARY_OUTPUT_RELOAD_CLASS}
2750 was supposed to be defined be defined to return the largest register
2751 class required. If the
2752 requirements for input and output reloads were the same, the macro
2753 @code{SECONDARY_RELOAD_CLASS} should have been used instead of defining both
2756 The values returned by these macros are often @code{GENERAL_REGS}.
2757 Return @code{NO_REGS} if no spare register is needed; i.e., if @var{x}
2758 can be directly copied to or from a register of @var{class} in
2759 @var{mode} without requiring a scratch register. Do not define this
2760 macro if it would always return @code{NO_REGS}.
2762 If a scratch register is required (either with or without an
2763 intermediate register), you were supposed to define patterns for
2764 @samp{reload_in@var{m}} or @samp{reload_out@var{m}}, as required
2765 (@pxref{Standard Names}. These patterns, which were normally
2766 implemented with a @code{define_expand}, should be similar to the
2767 @samp{mov@var{m}} patterns, except that operand 2 is the scratch
2770 These patterns need constraints for the reload register and scratch
2772 contain a single register class. If the original reload register (whose
2773 class is @var{class}) can meet the constraint given in the pattern, the
2774 value returned by these macros is used for the class of the scratch
2775 register. Otherwise, two additional reload registers are required.
2776 Their classes are obtained from the constraints in the insn pattern.
2778 @var{x} might be a pseudo-register or a @code{subreg} of a
2779 pseudo-register, which could either be in a hard register or in memory.
2780 Use @code{true_regnum} to find out; it will return @minus{}1 if the pseudo is
2781 in memory and the hard register number if it is in a register.
2783 These macros should not be used in the case where a particular class of
2784 registers can only be copied to memory and not to another class of
2785 registers. In that case, secondary reload registers are not needed and
2786 would not be helpful. Instead, a stack location must be used to perform
2787 the copy and the @code{mov@var{m}} pattern should use memory as an
2788 intermediate storage. This case often occurs between floating-point and
2792 @defmac SECONDARY_MEMORY_NEEDED (@var{class1}, @var{class2}, @var{m})
2793 Certain machines have the property that some registers cannot be copied
2794 to some other registers without using memory. Define this macro on
2795 those machines to be a C expression that is nonzero if objects of mode
2796 @var{m} in registers of @var{class1} can only be copied to registers of
2797 class @var{class2} by storing a register of @var{class1} into memory
2798 and loading that memory location into a register of @var{class2}.
2800 Do not define this macro if its value would always be zero.
2803 @defmac SECONDARY_MEMORY_NEEDED_RTX (@var{mode})
2804 Normally when @code{SECONDARY_MEMORY_NEEDED} is defined, the compiler
2805 allocates a stack slot for a memory location needed for register copies.
2806 If this macro is defined, the compiler instead uses the memory location
2807 defined by this macro.
2809 Do not define this macro if you do not define
2810 @code{SECONDARY_MEMORY_NEEDED}.
2813 @defmac SECONDARY_MEMORY_NEEDED_MODE (@var{mode})
2814 When the compiler needs a secondary memory location to copy between two
2815 registers of mode @var{mode}, it normally allocates sufficient memory to
2816 hold a quantity of @code{BITS_PER_WORD} bits and performs the store and
2817 load operations in a mode that many bits wide and whose class is the
2818 same as that of @var{mode}.
2820 This is right thing to do on most machines because it ensures that all
2821 bits of the register are copied and prevents accesses to the registers
2822 in a narrower mode, which some machines prohibit for floating-point
2825 However, this default behavior is not correct on some machines, such as
2826 the DEC Alpha, that store short integers in floating-point registers
2827 differently than in integer registers. On those machines, the default
2828 widening will not work correctly and you must define this macro to
2829 suppress that widening in some cases. See the file @file{alpha.h} for
2832 Do not define this macro if you do not define
2833 @code{SECONDARY_MEMORY_NEEDED} or if widening @var{mode} to a mode that
2834 is @code{BITS_PER_WORD} bits wide is correct for your machine.
2837 @deftypefn {Target Hook} bool TARGET_CLASS_LIKELY_SPILLED_P (reg_class_t @var{rclass})
2838 A target hook which returns @code{true} if pseudos that have been assigned
2839 to registers of class @var{rclass} would likely be spilled because
2840 registers of @var{rclass} are needed for spill registers.
2842 The default version of this target hook returns @code{true} if @var{rclass}
2843 has exactly one register and @code{false} otherwise. On most machines, this
2844 default should be used. Only use this target hook to some other expression
2845 if pseudos allocated by @file{local-alloc.c} end up in memory because their
2846 hard registers were needed for spill registers. If this target hook returns
2847 @code{false} for those classes, those pseudos will only be allocated by
2848 @file{global.c}, which knows how to reallocate the pseudo to another
2849 register. If there would not be another register available for reallocation,
2850 you should not change the implementation of this target hook since
2851 the only effect of such implementation would be to slow down register
2855 @defmac CLASS_LIKELY_SPILLED_P (@var{class})
2856 A C expression whose value is nonzero if pseudos that have been assigned
2857 to registers of class @var{class} would likely be spilled because
2858 registers of @var{class} are needed for spill registers.
2860 The default value of this macro returns 1 if @var{class} has exactly one
2861 register and zero otherwise. On most machines, this default should be
2862 used. Only define this macro to some other expression if pseudos
2863 allocated by @file{local-alloc.c} end up in memory because their hard
2864 registers were needed for spill registers. If this macro returns nonzero
2865 for those classes, those pseudos will only be allocated by
2866 @file{global.c}, which knows how to reallocate the pseudo to another
2867 register. If there would not be another register available for
2868 reallocation, you should not change the definition of this macro since
2869 the only effect of such a definition would be to slow down register
2873 @defmac CLASS_MAX_NREGS (@var{class}, @var{mode})
2874 A C expression for the maximum number of consecutive registers
2875 of class @var{class} needed to hold a value of mode @var{mode}.
2877 This is closely related to the macro @code{HARD_REGNO_NREGS}. In fact,
2878 the value of the macro @code{CLASS_MAX_NREGS (@var{class}, @var{mode})}
2879 should be the maximum value of @code{HARD_REGNO_NREGS (@var{regno},
2880 @var{mode})} for all @var{regno} values in the class @var{class}.
2882 This macro helps control the handling of multiple-word values
2886 @defmac CANNOT_CHANGE_MODE_CLASS (@var{from}, @var{to}, @var{class})
2887 If defined, a C expression that returns nonzero for a @var{class} for which
2888 a change from mode @var{from} to mode @var{to} is invalid.
2890 For the example, loading 32-bit integer or floating-point objects into
2891 floating-point registers on the Alpha extends them to 64 bits.
2892 Therefore loading a 64-bit object and then storing it as a 32-bit object
2893 does not store the low-order 32 bits, as would be the case for a normal
2894 register. Therefore, @file{alpha.h} defines @code{CANNOT_CHANGE_MODE_CLASS}
2898 #define CANNOT_CHANGE_MODE_CLASS(FROM, TO, CLASS) \
2899 (GET_MODE_SIZE (FROM) != GET_MODE_SIZE (TO) \
2900 ? reg_classes_intersect_p (FLOAT_REGS, (CLASS)) : 0)
2904 @deftypefn {Target Hook} {const reg_class_t *} TARGET_IRA_COVER_CLASSES (void)
2905 Return an array of cover classes for the Integrated Register Allocator
2906 (@acronym{IRA}). Cover classes are a set of non-intersecting register
2907 classes covering all hard registers used for register allocation
2908 purposes. If a move between two registers in the same cover class is
2909 possible, it should be cheaper than a load or store of the registers.
2910 The array is terminated by a @code{LIM_REG_CLASSES} element.
2912 The order of cover classes in the array is important. If two classes
2913 have the same cost of usage for a pseudo, the class occurred first in
2914 the array is chosen for the pseudo.
2916 This hook is called once at compiler startup, after the command-line
2917 options have been processed. It is then re-examined by every call to
2918 @code{target_reinit}.
2920 The default implementation returns @code{IRA_COVER_CLASSES}, if defined,
2921 otherwise there is no default implementation. You must define either this
2922 macro or @code{IRA_COVER_CLASSES} in order to use the integrated register
2923 allocator with Chaitin-Briggs coloring. If the macro is not defined,
2924 the only available coloring algorithm is Chow's priority coloring.
2927 @defmac IRA_COVER_CLASSES
2928 See the documentation for @code{TARGET_IRA_COVER_CLASSES}.
2931 @node Old Constraints
2932 @section Obsolete Macros for Defining Constraints
2933 @cindex defining constraints, obsolete method
2934 @cindex constraints, defining, obsolete method
2936 Machine-specific constraints can be defined with these macros instead
2937 of the machine description constructs described in @ref{Define
2938 Constraints}. This mechanism is obsolete. New ports should not use
2939 it; old ports should convert to the new mechanism.
2941 @defmac CONSTRAINT_LEN (@var{char}, @var{str})
2942 For the constraint at the start of @var{str}, which starts with the letter
2943 @var{c}, return the length. This allows you to have register class /
2944 constant / extra constraints that are longer than a single letter;
2945 you don't need to define this macro if you can do with single-letter
2946 constraints only. The definition of this macro should use
2947 DEFAULT_CONSTRAINT_LEN for all the characters that you don't want
2948 to handle specially.
2949 There are some sanity checks in genoutput.c that check the constraint lengths
2950 for the md file, so you can also use this macro to help you while you are
2951 transitioning from a byzantine single-letter-constraint scheme: when you
2952 return a negative length for a constraint you want to re-use, genoutput
2953 will complain about every instance where it is used in the md file.
2956 @defmac REG_CLASS_FROM_LETTER (@var{char})
2957 A C expression which defines the machine-dependent operand constraint
2958 letters for register classes. If @var{char} is such a letter, the
2959 value should be the register class corresponding to it. Otherwise,
2960 the value should be @code{NO_REGS}. The register letter @samp{r},
2961 corresponding to class @code{GENERAL_REGS}, will not be passed
2962 to this macro; you do not need to handle it.
2965 @defmac REG_CLASS_FROM_CONSTRAINT (@var{char}, @var{str})
2966 Like @code{REG_CLASS_FROM_LETTER}, but you also get the constraint string
2967 passed in @var{str}, so that you can use suffixes to distinguish between
2971 @defmac CONST_OK_FOR_LETTER_P (@var{value}, @var{c})
2972 A C expression that defines the machine-dependent operand constraint
2973 letters (@samp{I}, @samp{J}, @samp{K}, @dots{} @samp{P}) that specify
2974 particular ranges of integer values. If @var{c} is one of those
2975 letters, the expression should check that @var{value}, an integer, is in
2976 the appropriate range and return 1 if so, 0 otherwise. If @var{c} is
2977 not one of those letters, the value should be 0 regardless of
2981 @defmac CONST_OK_FOR_CONSTRAINT_P (@var{value}, @var{c}, @var{str})
2982 Like @code{CONST_OK_FOR_LETTER_P}, but you also get the constraint
2983 string passed in @var{str}, so that you can use suffixes to distinguish
2984 between different variants.
2987 @defmac CONST_DOUBLE_OK_FOR_LETTER_P (@var{value}, @var{c})
2988 A C expression that defines the machine-dependent operand constraint
2989 letters that specify particular ranges of @code{const_double} values
2990 (@samp{G} or @samp{H}).
2992 If @var{c} is one of those letters, the expression should check that
2993 @var{value}, an RTX of code @code{const_double}, is in the appropriate
2994 range and return 1 if so, 0 otherwise. If @var{c} is not one of those
2995 letters, the value should be 0 regardless of @var{value}.
2997 @code{const_double} is used for all floating-point constants and for
2998 @code{DImode} fixed-point constants. A given letter can accept either
2999 or both kinds of values. It can use @code{GET_MODE} to distinguish
3000 between these kinds.
3003 @defmac CONST_DOUBLE_OK_FOR_CONSTRAINT_P (@var{value}, @var{c}, @var{str})
3004 Like @code{CONST_DOUBLE_OK_FOR_LETTER_P}, but you also get the constraint
3005 string passed in @var{str}, so that you can use suffixes to distinguish
3006 between different variants.
3009 @defmac EXTRA_CONSTRAINT (@var{value}, @var{c})
3010 A C expression that defines the optional machine-dependent constraint
3011 letters that can be used to segregate specific types of operands, usually
3012 memory references, for the target machine. Any letter that is not
3013 elsewhere defined and not matched by @code{REG_CLASS_FROM_LETTER} /
3014 @code{REG_CLASS_FROM_CONSTRAINT}
3015 may be used. Normally this macro will not be defined.
3017 If it is required for a particular target machine, it should return 1
3018 if @var{value} corresponds to the operand type represented by the
3019 constraint letter @var{c}. If @var{c} is not defined as an extra
3020 constraint, the value returned should be 0 regardless of @var{value}.
3022 For example, on the ROMP, load instructions cannot have their output
3023 in r0 if the memory reference contains a symbolic address. Constraint
3024 letter @samp{Q} is defined as representing a memory address that does
3025 @emph{not} contain a symbolic address. An alternative is specified with
3026 a @samp{Q} constraint on the input and @samp{r} on the output. The next
3027 alternative specifies @samp{m} on the input and a register class that
3028 does not include r0 on the output.
3031 @defmac EXTRA_CONSTRAINT_STR (@var{value}, @var{c}, @var{str})
3032 Like @code{EXTRA_CONSTRAINT}, but you also get the constraint string passed
3033 in @var{str}, so that you can use suffixes to distinguish between different
3037 @defmac EXTRA_MEMORY_CONSTRAINT (@var{c}, @var{str})
3038 A C expression that defines the optional machine-dependent constraint
3039 letters, amongst those accepted by @code{EXTRA_CONSTRAINT}, that should
3040 be treated like memory constraints by the reload pass.
3042 It should return 1 if the operand type represented by the constraint
3043 at the start of @var{str}, the first letter of which is the letter @var{c},
3044 comprises a subset of all memory references including
3045 all those whose address is simply a base register. This allows the reload
3046 pass to reload an operand, if it does not directly correspond to the operand
3047 type of @var{c}, by copying its address into a base register.
3049 For example, on the S/390, some instructions do not accept arbitrary
3050 memory references, but only those that do not make use of an index
3051 register. The constraint letter @samp{Q} is defined via
3052 @code{EXTRA_CONSTRAINT} as representing a memory address of this type.
3053 If the letter @samp{Q} is marked as @code{EXTRA_MEMORY_CONSTRAINT},
3054 a @samp{Q} constraint can handle any memory operand, because the
3055 reload pass knows it can be reloaded by copying the memory address
3056 into a base register if required. This is analogous to the way
3057 an @samp{o} constraint can handle any memory operand.
3060 @defmac EXTRA_ADDRESS_CONSTRAINT (@var{c}, @var{str})
3061 A C expression that defines the optional machine-dependent constraint
3062 letters, amongst those accepted by @code{EXTRA_CONSTRAINT} /
3063 @code{EXTRA_CONSTRAINT_STR}, that should
3064 be treated like address constraints by the reload pass.
3066 It should return 1 if the operand type represented by the constraint
3067 at the start of @var{str}, which starts with the letter @var{c}, comprises
3068 a subset of all memory addresses including
3069 all those that consist of just a base register. This allows the reload
3070 pass to reload an operand, if it does not directly correspond to the operand
3071 type of @var{str}, by copying it into a base register.
3073 Any constraint marked as @code{EXTRA_ADDRESS_CONSTRAINT} can only
3074 be used with the @code{address_operand} predicate. It is treated
3075 analogously to the @samp{p} constraint.
3078 @node Stack and Calling
3079 @section Stack Layout and Calling Conventions
3080 @cindex calling conventions
3082 @c prevent bad page break with this line
3083 This describes the stack layout and calling conventions.
3087 * Exception Handling::
3092 * Register Arguments::
3094 * Aggregate Return::
3099 * Stack Smashing Protection::
3103 @subsection Basic Stack Layout
3104 @cindex stack frame layout
3105 @cindex frame layout
3107 @c prevent bad page break with this line
3108 Here is the basic stack layout.
3110 @defmac STACK_GROWS_DOWNWARD
3111 Define this macro if pushing a word onto the stack moves the stack
3112 pointer to a smaller address.
3114 When we say, ``define this macro if @dots{}'', it means that the
3115 compiler checks this macro only with @code{#ifdef} so the precise
3116 definition used does not matter.
3119 @defmac STACK_PUSH_CODE
3120 This macro defines the operation used when something is pushed
3121 on the stack. In RTL, a push operation will be
3122 @code{(set (mem (STACK_PUSH_CODE (reg sp))) @dots{})}
3124 The choices are @code{PRE_DEC}, @code{POST_DEC}, @code{PRE_INC},
3125 and @code{POST_INC}. Which of these is correct depends on
3126 the stack direction and on whether the stack pointer points
3127 to the last item on the stack or whether it points to the
3128 space for the next item on the stack.
3130 The default is @code{PRE_DEC} when @code{STACK_GROWS_DOWNWARD} is
3131 defined, which is almost always right, and @code{PRE_INC} otherwise,
3132 which is often wrong.
3135 @defmac FRAME_GROWS_DOWNWARD
3136 Define this macro to nonzero value if the addresses of local variable slots
3137 are at negative offsets from the frame pointer.
3140 @defmac ARGS_GROW_DOWNWARD
3141 Define this macro if successive arguments to a function occupy decreasing
3142 addresses on the stack.
3145 @defmac STARTING_FRAME_OFFSET
3146 Offset from the frame pointer to the first local variable slot to be allocated.
3148 If @code{FRAME_GROWS_DOWNWARD}, find the next slot's offset by
3149 subtracting the first slot's length from @code{STARTING_FRAME_OFFSET}.
3150 Otherwise, it is found by adding the length of the first slot to the
3151 value @code{STARTING_FRAME_OFFSET}.
3152 @c i'm not sure if the above is still correct.. had to change it to get
3153 @c rid of an overfull. --mew 2feb93
3156 @defmac STACK_ALIGNMENT_NEEDED
3157 Define to zero to disable final alignment of the stack during reload.
3158 The nonzero default for this macro is suitable for most ports.
3160 On ports where @code{STARTING_FRAME_OFFSET} is nonzero or where there
3161 is a register save block following the local block that doesn't require
3162 alignment to @code{STACK_BOUNDARY}, it may be beneficial to disable
3163 stack alignment and do it in the backend.
3166 @defmac STACK_POINTER_OFFSET
3167 Offset from the stack pointer register to the first location at which
3168 outgoing arguments are placed. If not specified, the default value of
3169 zero is used. This is the proper value for most machines.
3171 If @code{ARGS_GROW_DOWNWARD}, this is the offset to the location above
3172 the first location at which outgoing arguments are placed.
3175 @defmac FIRST_PARM_OFFSET (@var{fundecl})
3176 Offset from the argument pointer register to the first argument's
3177 address. On some machines it may depend on the data type of the
3180 If @code{ARGS_GROW_DOWNWARD}, this is the offset to the location above
3181 the first argument's address.
3184 @defmac STACK_DYNAMIC_OFFSET (@var{fundecl})
3185 Offset from the stack pointer register to an item dynamically allocated
3186 on the stack, e.g., by @code{alloca}.
3188 The default value for this macro is @code{STACK_POINTER_OFFSET} plus the
3189 length of the outgoing arguments. The default is correct for most
3190 machines. See @file{function.c} for details.
3193 @defmac INITIAL_FRAME_ADDRESS_RTX
3194 A C expression whose value is RTL representing the address of the initial
3195 stack frame. This address is passed to @code{RETURN_ADDR_RTX} and
3196 @code{DYNAMIC_CHAIN_ADDRESS}. If you don't define this macro, a reasonable
3197 default value will be used. Define this macro in order to make frame pointer
3198 elimination work in the presence of @code{__builtin_frame_address (count)} and
3199 @code{__builtin_return_address (count)} for @code{count} not equal to zero.
3202 @defmac DYNAMIC_CHAIN_ADDRESS (@var{frameaddr})
3203 A C expression whose value is RTL representing the address in a stack
3204 frame where the pointer to the caller's frame is stored. Assume that
3205 @var{frameaddr} is an RTL expression for the address of the stack frame
3208 If you don't define this macro, the default is to return the value
3209 of @var{frameaddr}---that is, the stack frame address is also the
3210 address of the stack word that points to the previous frame.
3213 @defmac SETUP_FRAME_ADDRESSES
3214 If defined, a C expression that produces the machine-specific code to
3215 setup the stack so that arbitrary frames can be accessed. For example,
3216 on the SPARC, we must flush all of the register windows to the stack
3217 before we can access arbitrary stack frames. You will seldom need to
3221 @deftypefn {Target Hook} rtx TARGET_BUILTIN_SETJMP_FRAME_VALUE (void)
3222 This target hook should return an rtx that is used to store
3223 the address of the current frame into the built in @code{setjmp} buffer.
3224 The default value, @code{virtual_stack_vars_rtx}, is correct for most
3225 machines. One reason you may need to define this target hook is if
3226 @code{hard_frame_pointer_rtx} is the appropriate value on your machine.
3229 @defmac FRAME_ADDR_RTX (@var{frameaddr})
3230 A C expression whose value is RTL representing the value of the frame
3231 address for the current frame. @var{frameaddr} is the frame pointer
3232 of the current frame. This is used for __builtin_frame_address.
3233 You need only define this macro if the frame address is not the same
3234 as the frame pointer. Most machines do not need to define it.
3237 @defmac RETURN_ADDR_RTX (@var{count}, @var{frameaddr})
3238 A C expression whose value is RTL representing the value of the return
3239 address for the frame @var{count} steps up from the current frame, after
3240 the prologue. @var{frameaddr} is the frame pointer of the @var{count}
3241 frame, or the frame pointer of the @var{count} @minus{} 1 frame if
3242 @code{RETURN_ADDR_IN_PREVIOUS_FRAME} is defined.
3244 The value of the expression must always be the correct address when
3245 @var{count} is zero, but may be @code{NULL_RTX} if there is no way to
3246 determine the return address of other frames.
3249 @defmac RETURN_ADDR_IN_PREVIOUS_FRAME
3250 Define this if the return address of a particular stack frame is accessed
3251 from the frame pointer of the previous stack frame.
3254 @defmac INCOMING_RETURN_ADDR_RTX
3255 A C expression whose value is RTL representing the location of the
3256 incoming return address at the beginning of any function, before the
3257 prologue. This RTL is either a @code{REG}, indicating that the return
3258 value is saved in @samp{REG}, or a @code{MEM} representing a location in
3261 You only need to define this macro if you want to support call frame
3262 debugging information like that provided by DWARF 2.
3264 If this RTL is a @code{REG}, you should also define
3265 @code{DWARF_FRAME_RETURN_COLUMN} to @code{DWARF_FRAME_REGNUM (REGNO)}.
3268 @defmac DWARF_ALT_FRAME_RETURN_COLUMN
3269 A C expression whose value is an integer giving a DWARF 2 column
3270 number that may be used as an alternative return column. The column
3271 must not correspond to any gcc hard register (that is, it must not
3272 be in the range of @code{DWARF_FRAME_REGNUM}).
3274 This macro can be useful if @code{DWARF_FRAME_RETURN_COLUMN} is set to a
3275 general register, but an alternative column needs to be used for signal
3276 frames. Some targets have also used different frame return columns
3280 @defmac DWARF_ZERO_REG
3281 A C expression whose value is an integer giving a DWARF 2 register
3282 number that is considered to always have the value zero. This should
3283 only be defined if the target has an architected zero register, and
3284 someone decided it was a good idea to use that register number to
3285 terminate the stack backtrace. New ports should avoid this.
3288 @deftypefn {Target Hook} void TARGET_DWARF_HANDLE_FRAME_UNSPEC (const char *@var{label}, rtx @var{pattern}, int @var{index})
3289 This target hook allows the backend to emit frame-related insns that
3290 contain UNSPECs or UNSPEC_VOLATILEs. The DWARF 2 call frame debugging
3291 info engine will invoke it on insns of the form
3293 (set (reg) (unspec [@dots{}] UNSPEC_INDEX))
3297 (set (reg) (unspec_volatile [@dots{}] UNSPECV_INDEX)).
3299 to let the backend emit the call frame instructions. @var{label} is
3300 the CFI label attached to the insn, @var{pattern} is the pattern of
3301 the insn and @var{index} is @code{UNSPEC_INDEX} or @code{UNSPECV_INDEX}.
3304 @defmac INCOMING_FRAME_SP_OFFSET
3305 A C expression whose value is an integer giving the offset, in bytes,
3306 from the value of the stack pointer register to the top of the stack
3307 frame at the beginning of any function, before the prologue. The top of
3308 the frame is defined to be the value of the stack pointer in the
3309 previous frame, just before the call instruction.
3311 You only need to define this macro if you want to support call frame
3312 debugging information like that provided by DWARF 2.
3315 @defmac ARG_POINTER_CFA_OFFSET (@var{fundecl})
3316 A C expression whose value is an integer giving the offset, in bytes,
3317 from the argument pointer to the canonical frame address (cfa). The
3318 final value should coincide with that calculated by
3319 @code{INCOMING_FRAME_SP_OFFSET}. Which is unfortunately not usable
3320 during virtual register instantiation.
3322 The default value for this macro is
3323 @code{FIRST_PARM_OFFSET (fundecl) + crtl->args.pretend_args_size},
3324 which is correct for most machines; in general, the arguments are found
3325 immediately before the stack frame. Note that this is not the case on
3326 some targets that save registers into the caller's frame, such as SPARC
3327 and rs6000, and so such targets need to define this macro.
3329 You only need to define this macro if the default is incorrect, and you
3330 want to support call frame debugging information like that provided by
3334 @defmac FRAME_POINTER_CFA_OFFSET (@var{fundecl})
3335 If defined, a C expression whose value is an integer giving the offset
3336 in bytes from the frame pointer to the canonical frame address (cfa).
3337 The final value should coincide with that calculated by
3338 @code{INCOMING_FRAME_SP_OFFSET}.
3340 Normally the CFA is calculated as an offset from the argument pointer,
3341 via @code{ARG_POINTER_CFA_OFFSET}, but if the argument pointer is
3342 variable due to the ABI, this may not be possible. If this macro is
3343 defined, it implies that the virtual register instantiation should be
3344 based on the frame pointer instead of the argument pointer. Only one
3345 of @code{FRAME_POINTER_CFA_OFFSET} and @code{ARG_POINTER_CFA_OFFSET}
3349 @defmac CFA_FRAME_BASE_OFFSET (@var{fundecl})
3350 If defined, a C expression whose value is an integer giving the offset
3351 in bytes from the canonical frame address (cfa) to the frame base used
3352 in DWARF 2 debug information. The default is zero. A different value
3353 may reduce the size of debug information on some ports.
3356 @node Exception Handling
3357 @subsection Exception Handling Support
3358 @cindex exception handling
3360 @defmac EH_RETURN_DATA_REGNO (@var{N})
3361 A C expression whose value is the @var{N}th register number used for
3362 data by exception handlers, or @code{INVALID_REGNUM} if fewer than
3363 @var{N} registers are usable.
3365 The exception handling library routines communicate with the exception
3366 handlers via a set of agreed upon registers. Ideally these registers
3367 should be call-clobbered; it is possible to use call-saved registers,
3368 but may negatively impact code size. The target must support at least
3369 2 data registers, but should define 4 if there are enough free registers.
3371 You must define this macro if you want to support call frame exception
3372 handling like that provided by DWARF 2.
3375 @defmac EH_RETURN_STACKADJ_RTX
3376 A C expression whose value is RTL representing a location in which
3377 to store a stack adjustment to be applied before function return.
3378 This is used to unwind the stack to an exception handler's call frame.
3379 It will be assigned zero on code paths that return normally.
3381 Typically this is a call-clobbered hard register that is otherwise
3382 untouched by the epilogue, but could also be a stack slot.
3384 Do not define this macro if the stack pointer is saved and restored
3385 by the regular prolog and epilog code in the call frame itself; in
3386 this case, the exception handling library routines will update the
3387 stack location to be restored in place. Otherwise, you must define
3388 this macro if you want to support call frame exception handling like
3389 that provided by DWARF 2.
3392 @defmac EH_RETURN_HANDLER_RTX
3393 A C expression whose value is RTL representing a location in which
3394 to store the address of an exception handler to which we should
3395 return. It will not be assigned on code paths that return normally.
3397 Typically this is the location in the call frame at which the normal
3398 return address is stored. For targets that return by popping an
3399 address off the stack, this might be a memory address just below
3400 the @emph{target} call frame rather than inside the current call
3401 frame. If defined, @code{EH_RETURN_STACKADJ_RTX} will have already
3402 been assigned, so it may be used to calculate the location of the
3405 Some targets have more complex requirements than storing to an
3406 address calculable during initial code generation. In that case
3407 the @code{eh_return} instruction pattern should be used instead.
3409 If you want to support call frame exception handling, you must
3410 define either this macro or the @code{eh_return} instruction pattern.
3413 @defmac RETURN_ADDR_OFFSET
3414 If defined, an integer-valued C expression for which rtl will be generated
3415 to add it to the exception handler address before it is searched in the
3416 exception handling tables, and to subtract it again from the address before
3417 using it to return to the exception handler.
3420 @defmac ASM_PREFERRED_EH_DATA_FORMAT (@var{code}, @var{global})
3421 This macro chooses the encoding of pointers embedded in the exception
3422 handling sections. If at all possible, this should be defined such
3423 that the exception handling section will not require dynamic relocations,
3424 and so may be read-only.
3426 @var{code} is 0 for data, 1 for code labels, 2 for function pointers.
3427 @var{global} is true if the symbol may be affected by dynamic relocations.
3428 The macro should return a combination of the @code{DW_EH_PE_*} defines
3429 as found in @file{dwarf2.h}.
3431 If this macro is not defined, pointers will not be encoded but
3432 represented directly.
3435 @defmac ASM_MAYBE_OUTPUT_ENCODED_ADDR_RTX (@var{file}, @var{encoding}, @var{size}, @var{addr}, @var{done})
3436 This macro allows the target to emit whatever special magic is required
3437 to represent the encoding chosen by @code{ASM_PREFERRED_EH_DATA_FORMAT}.
3438 Generic code takes care of pc-relative and indirect encodings; this must
3439 be defined if the target uses text-relative or data-relative encodings.
3441 This is a C statement that branches to @var{done} if the format was
3442 handled. @var{encoding} is the format chosen, @var{size} is the number
3443 of bytes that the format occupies, @var{addr} is the @code{SYMBOL_REF}
3447 @defmac MD_UNWIND_SUPPORT
3448 A string specifying a file to be #include'd in unwind-dw2.c. The file
3449 so included typically defines @code{MD_FALLBACK_FRAME_STATE_FOR}.
3452 @defmac MD_FALLBACK_FRAME_STATE_FOR (@var{context}, @var{fs})
3453 This macro allows the target to add CPU and operating system specific
3454 code to the call-frame unwinder for use when there is no unwind data
3455 available. The most common reason to implement this macro is to unwind
3456 through signal frames.
3458 This macro is called from @code{uw_frame_state_for} in
3459 @file{unwind-dw2.c}, @file{unwind-dw2-xtensa.c} and
3460 @file{unwind-ia64.c}. @var{context} is an @code{_Unwind_Context};
3461 @var{fs} is an @code{_Unwind_FrameState}. Examine @code{context->ra}
3462 for the address of the code being executed and @code{context->cfa} for
3463 the stack pointer value. If the frame can be decoded, the register
3464 save addresses should be updated in @var{fs} and the macro should
3465 evaluate to @code{_URC_NO_REASON}. If the frame cannot be decoded,
3466 the macro should evaluate to @code{_URC_END_OF_STACK}.
3468 For proper signal handling in Java this macro is accompanied by
3469 @code{MAKE_THROW_FRAME}, defined in @file{libjava/include/*-signal.h} headers.
3472 @defmac MD_HANDLE_UNWABI (@var{context}, @var{fs})
3473 This macro allows the target to add operating system specific code to the
3474 call-frame unwinder to handle the IA-64 @code{.unwabi} unwinding directive,
3475 usually used for signal or interrupt frames.
3477 This macro is called from @code{uw_update_context} in @file{unwind-ia64.c}.
3478 @var{context} is an @code{_Unwind_Context};
3479 @var{fs} is an @code{_Unwind_FrameState}. Examine @code{fs->unwabi}
3480 for the abi and context in the @code{.unwabi} directive. If the
3481 @code{.unwabi} directive can be handled, the register save addresses should
3482 be updated in @var{fs}.
3485 @defmac TARGET_USES_WEAK_UNWIND_INFO
3486 A C expression that evaluates to true if the target requires unwind
3487 info to be given comdat linkage. Define it to be @code{1} if comdat
3488 linkage is necessary. The default is @code{0}.
3491 @node Stack Checking
3492 @subsection Specifying How Stack Checking is Done
3494 GCC will check that stack references are within the boundaries of the
3495 stack, if the option @option{-fstack-check} is specified, in one of
3500 If the value of the @code{STACK_CHECK_BUILTIN} macro is nonzero, GCC
3501 will assume that you have arranged for full stack checking to be done
3502 at appropriate places in the configuration files. GCC will not do
3503 other special processing.
3506 If @code{STACK_CHECK_BUILTIN} is zero and the value of the
3507 @code{STACK_CHECK_STATIC_BUILTIN} macro is nonzero, GCC will assume
3508 that you have arranged for static stack checking (checking of the
3509 static stack frame of functions) to be done at appropriate places
3510 in the configuration files. GCC will only emit code to do dynamic
3511 stack checking (checking on dynamic stack allocations) using the third
3515 If neither of the above are true, GCC will generate code to periodically
3516 ``probe'' the stack pointer using the values of the macros defined below.
3519 If neither STACK_CHECK_BUILTIN nor STACK_CHECK_STATIC_BUILTIN is defined,
3520 GCC will change its allocation strategy for large objects if the option
3521 @option{-fstack-check} is specified: they will always be allocated
3522 dynamically if their size exceeds @code{STACK_CHECK_MAX_VAR_SIZE} bytes.
3524 @defmac STACK_CHECK_BUILTIN
3525 A nonzero value if stack checking is done by the configuration files in a
3526 machine-dependent manner. You should define this macro if stack checking
3527 is required by the ABI of your machine or if you would like to do stack
3528 checking in some more efficient way than the generic approach. The default
3529 value of this macro is zero.
3532 @defmac STACK_CHECK_STATIC_BUILTIN
3533 A nonzero value if static stack checking is done by the configuration files
3534 in a machine-dependent manner. You should define this macro if you would
3535 like to do static stack checking in some more efficient way than the generic
3536 approach. The default value of this macro is zero.
3539 @defmac STACK_CHECK_PROBE_INTERVAL_EXP
3540 An integer specifying the interval at which GCC must generate stack probe
3541 instructions, defined as 2 raised to this integer. You will normally
3542 define this macro so that the interval be no larger than the size of
3543 the ``guard pages'' at the end of a stack area. The default value
3544 of 12 (4096-byte interval) is suitable for most systems.
3547 @defmac STACK_CHECK_MOVING_SP
3548 An integer which is nonzero if GCC should move the stack pointer page by page
3549 when doing probes. This can be necessary on systems where the stack pointer
3550 contains the bottom address of the memory area accessible to the executing
3551 thread at any point in time. In this situation an alternate signal stack
3552 is required in order to be able to recover from a stack overflow. The
3553 default value of this macro is zero.
3556 @defmac STACK_CHECK_PROTECT
3557 The number of bytes of stack needed to recover from a stack overflow, for
3558 languages where such a recovery is supported. The default value of 75 words
3559 with the @code{setjmp}/@code{longjmp}-based exception handling mechanism and
3560 8192 bytes with other exception handling mechanisms should be adequate for
3564 The following macros are relevant only if neither STACK_CHECK_BUILTIN
3565 nor STACK_CHECK_STATIC_BUILTIN is defined; you can omit them altogether
3566 in the opposite case.
3568 @defmac STACK_CHECK_MAX_FRAME_SIZE
3569 The maximum size of a stack frame, in bytes. GCC will generate probe
3570 instructions in non-leaf functions to ensure at least this many bytes of
3571 stack are available. If a stack frame is larger than this size, stack
3572 checking will not be reliable and GCC will issue a warning. The
3573 default is chosen so that GCC only generates one instruction on most
3574 systems. You should normally not change the default value of this macro.
3577 @defmac STACK_CHECK_FIXED_FRAME_SIZE
3578 GCC uses this value to generate the above warning message. It
3579 represents the amount of fixed frame used by a function, not including
3580 space for any callee-saved registers, temporaries and user variables.
3581 You need only specify an upper bound for this amount and will normally
3582 use the default of four words.
3585 @defmac STACK_CHECK_MAX_VAR_SIZE
3586 The maximum size, in bytes, of an object that GCC will place in the
3587 fixed area of the stack frame when the user specifies
3588 @option{-fstack-check}.
3589 GCC computed the default from the values of the above macros and you will
3590 normally not need to override that default.
3594 @node Frame Registers
3595 @subsection Registers That Address the Stack Frame
3597 @c prevent bad page break with this line
3598 This discusses registers that address the stack frame.
3600 @defmac STACK_POINTER_REGNUM
3601 The register number of the stack pointer register, which must also be a
3602 fixed register according to @code{FIXED_REGISTERS}. On most machines,
3603 the hardware determines which register this is.
3606 @defmac FRAME_POINTER_REGNUM
3607 The register number of the frame pointer register, which is used to
3608 access automatic variables in the stack frame. On some machines, the
3609 hardware determines which register this is. On other machines, you can
3610 choose any register you wish for this purpose.
3613 @defmac HARD_FRAME_POINTER_REGNUM
3614 On some machines the offset between the frame pointer and starting
3615 offset of the automatic variables is not known until after register
3616 allocation has been done (for example, because the saved registers are
3617 between these two locations). On those machines, define
3618 @code{FRAME_POINTER_REGNUM} the number of a special, fixed register to
3619 be used internally until the offset is known, and define
3620 @code{HARD_FRAME_POINTER_REGNUM} to be the actual hard register number
3621 used for the frame pointer.
3623 You should define this macro only in the very rare circumstances when it
3624 is not possible to calculate the offset between the frame pointer and
3625 the automatic variables until after register allocation has been
3626 completed. When this macro is defined, you must also indicate in your
3627 definition of @code{ELIMINABLE_REGS} how to eliminate
3628 @code{FRAME_POINTER_REGNUM} into either @code{HARD_FRAME_POINTER_REGNUM}
3629 or @code{STACK_POINTER_REGNUM}.
3631 Do not define this macro if it would be the same as
3632 @code{FRAME_POINTER_REGNUM}.
3635 @defmac ARG_POINTER_REGNUM
3636 The register number of the arg pointer register, which is used to access
3637 the function's argument list. On some machines, this is the same as the
3638 frame pointer register. On some machines, the hardware determines which
3639 register this is. On other machines, you can choose any register you
3640 wish for this purpose. If this is not the same register as the frame
3641 pointer register, then you must mark it as a fixed register according to
3642 @code{FIXED_REGISTERS}, or arrange to be able to eliminate it
3643 (@pxref{Elimination}).
3646 @defmac RETURN_ADDRESS_POINTER_REGNUM
3647 The register number of the return address pointer register, which is used to
3648 access the current function's return address from the stack. On some
3649 machines, the return address is not at a fixed offset from the frame
3650 pointer or stack pointer or argument pointer. This register can be defined
3651 to point to the return address on the stack, and then be converted by
3652 @code{ELIMINABLE_REGS} into either the frame pointer or stack pointer.
3654 Do not define this macro unless there is no other way to get the return
3655 address from the stack.
3658 @defmac STATIC_CHAIN_REGNUM
3659 @defmacx STATIC_CHAIN_INCOMING_REGNUM
3660 Register numbers used for passing a function's static chain pointer. If
3661 register windows are used, the register number as seen by the called
3662 function is @code{STATIC_CHAIN_INCOMING_REGNUM}, while the register
3663 number as seen by the calling function is @code{STATIC_CHAIN_REGNUM}. If
3664 these registers are the same, @code{STATIC_CHAIN_INCOMING_REGNUM} need
3667 The static chain register need not be a fixed register.
3669 If the static chain is passed in memory, these macros should not be
3670 defined; instead, the @code{TARGET_STATIC_CHAIN} hook should be used.
3673 @deftypefn {Target Hook} rtx TARGET_STATIC_CHAIN (const_tree @var{fndecl}, bool @var{incoming_p})
3674 This hook replaces the use of @code{STATIC_CHAIN_REGNUM} et al for
3675 targets that may use different static chain locations for different
3676 nested functions. This may be required if the target has function
3677 attributes that affect the calling conventions of the function and
3678 those calling conventions use different static chain locations.
3680 The default version of this hook uses @code{STATIC_CHAIN_REGNUM} et al.
3682 If the static chain is passed in memory, this hook should be used to
3683 provide rtx giving @code{mem} expressions that denote where they are stored.
3684 Often the @code{mem} expression as seen by the caller will be at an offset
3685 from the stack pointer and the @code{mem} expression as seen by the callee
3686 will be at an offset from the frame pointer.
3687 @findex stack_pointer_rtx
3688 @findex frame_pointer_rtx
3689 @findex arg_pointer_rtx
3690 The variables @code{stack_pointer_rtx}, @code{frame_pointer_rtx}, and
3691 @code{arg_pointer_rtx} will have been initialized and should be used
3692 to refer to those items.
3695 @defmac DWARF_FRAME_REGISTERS
3696 This macro specifies the maximum number of hard registers that can be
3697 saved in a call frame. This is used to size data structures used in
3698 DWARF2 exception handling.
3700 Prior to GCC 3.0, this macro was needed in order to establish a stable
3701 exception handling ABI in the face of adding new hard registers for ISA
3702 extensions. In GCC 3.0 and later, the EH ABI is insulated from changes
3703 in the number of hard registers. Nevertheless, this macro can still be
3704 used to reduce the runtime memory requirements of the exception handling
3705 routines, which can be substantial if the ISA contains a lot of
3706 registers that are not call-saved.
3708 If this macro is not defined, it defaults to
3709 @code{FIRST_PSEUDO_REGISTER}.
3712 @defmac PRE_GCC3_DWARF_FRAME_REGISTERS
3714 This macro is similar to @code{DWARF_FRAME_REGISTERS}, but is provided
3715 for backward compatibility in pre GCC 3.0 compiled code.
3717 If this macro is not defined, it defaults to
3718 @code{DWARF_FRAME_REGISTERS}.
3721 @defmac DWARF_REG_TO_UNWIND_COLUMN (@var{regno})
3723 Define this macro if the target's representation for dwarf registers
3724 is different than the internal representation for unwind column.
3725 Given a dwarf register, this macro should return the internal unwind
3726 column number to use instead.
3728 See the PowerPC's SPE target for an example.
3731 @defmac DWARF_FRAME_REGNUM (@var{regno})
3733 Define this macro if the target's representation for dwarf registers
3734 used in .eh_frame or .debug_frame is different from that used in other
3735 debug info sections. Given a GCC hard register number, this macro
3736 should return the .eh_frame register number. The default is
3737 @code{DBX_REGISTER_NUMBER (@var{regno})}.
3741 @defmac DWARF2_FRAME_REG_OUT (@var{regno}, @var{for_eh})
3743 Define this macro to map register numbers held in the call frame info
3744 that GCC has collected using @code{DWARF_FRAME_REGNUM} to those that
3745 should be output in .debug_frame (@code{@var{for_eh}} is zero) and
3746 .eh_frame (@code{@var{for_eh}} is nonzero). The default is to
3747 return @code{@var{regno}}.
3752 @subsection Eliminating Frame Pointer and Arg Pointer
3754 @c prevent bad page break with this line
3755 This is about eliminating the frame pointer and arg pointer.
3757 @deftypefn {Target Hook} bool TARGET_FRAME_POINTER_REQUIRED (void)
3758 This target hook should return @code{true} if a function must have and use
3759 a frame pointer. This target hook is called in the reload pass. If its return
3760 value is @code{true} the function will have a frame pointer.
3762 This target hook can in principle examine the current function and decide
3763 according to the facts, but on most machines the constant @code{false} or the
3764 constant @code{true} suffices. Use @code{false} when the machine allows code
3765 to be generated with no frame pointer, and doing so saves some time or space.
3766 Use @code{true} when there is no possible advantage to avoiding a frame
3769 In certain cases, the compiler does not know how to produce valid code
3770 without a frame pointer. The compiler recognizes those cases and
3771 automatically gives the function a frame pointer regardless of what
3772 @code{TARGET_FRAME_POINTER_REQUIRED} returns. You don't need to worry about
3775 In a function that does not require a frame pointer, the frame pointer
3776 register can be allocated for ordinary usage, unless you mark it as a
3777 fixed register. See @code{FIXED_REGISTERS} for more information.
3779 Default return value is @code{false}.
3782 @findex get_frame_size
3783 @defmac INITIAL_FRAME_POINTER_OFFSET (@var{depth-var})
3784 A C statement to store in the variable @var{depth-var} the difference
3785 between the frame pointer and the stack pointer values immediately after
3786 the function prologue. The value would be computed from information
3787 such as the result of @code{get_frame_size ()} and the tables of
3788 registers @code{regs_ever_live} and @code{call_used_regs}.
3790 If @code{ELIMINABLE_REGS} is defined, this macro will be not be used and
3791 need not be defined. Otherwise, it must be defined even if
3792 @code{TARGET_FRAME_POINTER_REQUIRED} always returns true; in that
3793 case, you may set @var{depth-var} to anything.
3796 @defmac ELIMINABLE_REGS
3797 If defined, this macro specifies a table of register pairs used to
3798 eliminate unneeded registers that point into the stack frame. If it is not
3799 defined, the only elimination attempted by the compiler is to replace
3800 references to the frame pointer with references to the stack pointer.
3802 The definition of this macro is a list of structure initializations, each
3803 of which specifies an original and replacement register.
3805 On some machines, the position of the argument pointer is not known until
3806 the compilation is completed. In such a case, a separate hard register
3807 must be used for the argument pointer. This register can be eliminated by
3808 replacing it with either the frame pointer or the argument pointer,
3809 depending on whether or not the frame pointer has been eliminated.
3811 In this case, you might specify:
3813 #define ELIMINABLE_REGS \
3814 @{@{ARG_POINTER_REGNUM, STACK_POINTER_REGNUM@}, \
3815 @{ARG_POINTER_REGNUM, FRAME_POINTER_REGNUM@}, \
3816 @{FRAME_POINTER_REGNUM, STACK_POINTER_REGNUM@}@}
3819 Note that the elimination of the argument pointer with the stack pointer is
3820 specified first since that is the preferred elimination.
3823 @deftypefn {Target Hook} bool TARGET_CAN_ELIMINATE (const int @var{from_reg}, const int @var{to_reg})
3824 This target hook should returns @code{true} if the compiler is allowed to
3825 try to replace register number @var{from_reg} with register number
3826 @var{to_reg}. This target hook need only be defined if @code{ELIMINABLE_REGS}
3827 is defined, and will usually be @code{true}, since most of the cases
3828 preventing register elimination are things that the compiler already
3831 Default return value is @code{true}.
3834 @defmac INITIAL_ELIMINATION_OFFSET (@var{from-reg}, @var{to-reg}, @var{offset-var})
3835 This macro is similar to @code{INITIAL_FRAME_POINTER_OFFSET}. It
3836 specifies the initial difference between the specified pair of
3837 registers. This macro must be defined if @code{ELIMINABLE_REGS} is
3841 @node Stack Arguments
3842 @subsection Passing Function Arguments on the Stack
3843 @cindex arguments on stack
3844 @cindex stack arguments
3846 The macros in this section control how arguments are passed
3847 on the stack. See the following section for other macros that
3848 control passing certain arguments in registers.
3850 @deftypefn {Target Hook} bool TARGET_PROMOTE_PROTOTYPES (const_tree @var{fntype})
3851 This target hook returns @code{true} if an argument declared in a
3852 prototype as an integral type smaller than @code{int} should actually be
3853 passed as an @code{int}. In addition to avoiding errors in certain
3854 cases of mismatch, it also makes for better code on certain machines.
3855 The default is to not promote prototypes.
3859 A C expression. If nonzero, push insns will be used to pass
3861 If the target machine does not have a push instruction, set it to zero.
3862 That directs GCC to use an alternate strategy: to
3863 allocate the entire argument block and then store the arguments into
3864 it. When @code{PUSH_ARGS} is nonzero, @code{PUSH_ROUNDING} must be defined too.
3867 @defmac PUSH_ARGS_REVERSED
3868 A C expression. If nonzero, function arguments will be evaluated from
3869 last to first, rather than from first to last. If this macro is not
3870 defined, it defaults to @code{PUSH_ARGS} on targets where the stack
3871 and args grow in opposite directions, and 0 otherwise.
3874 @defmac PUSH_ROUNDING (@var{npushed})
3875 A C expression that is the number of bytes actually pushed onto the
3876 stack when an instruction attempts to push @var{npushed} bytes.
3878 On some machines, the definition
3881 #define PUSH_ROUNDING(BYTES) (BYTES)
3885 will suffice. But on other machines, instructions that appear
3886 to push one byte actually push two bytes in an attempt to maintain
3887 alignment. Then the definition should be
3890 #define PUSH_ROUNDING(BYTES) (((BYTES) + 1) & ~1)
3894 @findex current_function_outgoing_args_size
3895 @defmac ACCUMULATE_OUTGOING_ARGS
3896 A C expression. If nonzero, the maximum amount of space required for outgoing arguments
3897 will be computed and placed into the variable
3898 @code{current_function_outgoing_args_size}. No space will be pushed
3899 onto the stack for each call; instead, the function prologue should
3900 increase the stack frame size by this amount.
3902 Setting both @code{PUSH_ARGS} and @code{ACCUMULATE_OUTGOING_ARGS}
3906 @defmac REG_PARM_STACK_SPACE (@var{fndecl})
3907 Define this macro if functions should assume that stack space has been
3908 allocated for arguments even when their values are passed in
3911 The value of this macro is the size, in bytes, of the area reserved for
3912 arguments passed in registers for the function represented by @var{fndecl},
3913 which can be zero if GCC is calling a library function.
3914 The argument @var{fndecl} can be the FUNCTION_DECL, or the type itself
3917 This space can be allocated by the caller, or be a part of the
3918 machine-dependent stack frame: @code{OUTGOING_REG_PARM_STACK_SPACE} says
3921 @c above is overfull. not sure what to do. --mew 5feb93 did
3922 @c something, not sure if it looks good. --mew 10feb93
3924 @defmac OUTGOING_REG_PARM_STACK_SPACE (@var{fntype})
3925 Define this to a nonzero value if it is the responsibility of the
3926 caller to allocate the area reserved for arguments passed in registers
3927 when calling a function of @var{fntype}. @var{fntype} may be NULL
3928 if the function called is a library function.
3930 If @code{ACCUMULATE_OUTGOING_ARGS} is defined, this macro controls
3931 whether the space for these arguments counts in the value of
3932 @code{current_function_outgoing_args_size}.
3935 @defmac STACK_PARMS_IN_REG_PARM_AREA
3936 Define this macro if @code{REG_PARM_STACK_SPACE} is defined, but the
3937 stack parameters don't skip the area specified by it.
3938 @c i changed this, makes more sens and it should have taken care of the
3939 @c overfull.. not as specific, tho. --mew 5feb93
3941 Normally, when a parameter is not passed in registers, it is placed on the
3942 stack beyond the @code{REG_PARM_STACK_SPACE} area. Defining this macro
3943 suppresses this behavior and causes the parameter to be passed on the
3944 stack in its natural location.
3947 @deftypefn {Target Hook} int TARGET_RETURN_POPS_ARGS (tree @var{fundecl}, tree @var{funtype}, int @var{size})
3948 This target hook returns the number of bytes of its own arguments that
3949 a function pops on returning, or 0 if the function pops no arguments
3950 and the caller must therefore pop them all after the function returns.
3952 @var{fundecl} is a C variable whose value is a tree node that describes
3953 the function in question. Normally it is a node of type
3954 @code{FUNCTION_DECL} that describes the declaration of the function.
3955 From this you can obtain the @code{DECL_ATTRIBUTES} of the function.
3957 @var{funtype} is a C variable whose value is a tree node that
3958 describes the function in question. Normally it is a node of type
3959 @code{FUNCTION_TYPE} that describes the data type of the function.
3960 From this it is possible to obtain the data types of the value and
3961 arguments (if known).
3963 When a call to a library function is being considered, @var{fundecl}
3964 will contain an identifier node for the library function. Thus, if
3965 you need to distinguish among various library functions, you can do so
3966 by their names. Note that ``library function'' in this context means
3967 a function used to perform arithmetic, whose name is known specially
3968 in the compiler and was not mentioned in the C code being compiled.
3970 @var{size} is the number of bytes of arguments passed on the
3971 stack. If a variable number of bytes is passed, it is zero, and
3972 argument popping will always be the responsibility of the calling function.
3974 On the VAX, all functions always pop their arguments, so the definition
3975 of this macro is @var{size}. On the 68000, using the standard
3976 calling convention, no functions pop their arguments, so the value of
3977 the macro is always 0 in this case. But an alternative calling
3978 convention is available in which functions that take a fixed number of
3979 arguments pop them but other functions (such as @code{printf}) pop
3980 nothing (the caller pops all). When this convention is in use,
3981 @var{funtype} is examined to determine whether a function takes a fixed
3982 number of arguments.
3985 @defmac CALL_POPS_ARGS (@var{cum})
3986 A C expression that should indicate the number of bytes a call sequence
3987 pops off the stack. It is added to the value of @code{RETURN_POPS_ARGS}
3988 when compiling a function call.
3990 @var{cum} is the variable in which all arguments to the called function
3991 have been accumulated.
3993 On certain architectures, such as the SH5, a call trampoline is used
3994 that pops certain registers off the stack, depending on the arguments
3995 that have been passed to the function. Since this is a property of the
3996 call site, not of the called function, @code{RETURN_POPS_ARGS} is not
4000 @node Register Arguments
4001 @subsection Passing Arguments in Registers
4002 @cindex arguments in registers
4003 @cindex registers arguments
4005 This section describes the macros which let you control how various
4006 types of arguments are passed in registers or how they are arranged in
4009 @defmac FUNCTION_ARG (@var{cum}, @var{mode}, @var{type}, @var{named})
4010 A C expression that controls whether a function argument is passed
4011 in a register, and which register.
4013 The arguments are @var{cum}, which summarizes all the previous
4014 arguments; @var{mode}, the machine mode of the argument; @var{type},
4015 the data type of the argument as a tree node or 0 if that is not known
4016 (which happens for C support library functions); and @var{named},
4017 which is 1 for an ordinary argument and 0 for nameless arguments that
4018 correspond to @samp{@dots{}} in the called function's prototype.
4019 @var{type} can be an incomplete type if a syntax error has previously
4022 The value of the expression is usually either a @code{reg} RTX for the
4023 hard register in which to pass the argument, or zero to pass the
4024 argument on the stack.
4026 For machines like the VAX and 68000, where normally all arguments are
4027 pushed, zero suffices as a definition.
4029 The value of the expression can also be a @code{parallel} RTX@. This is
4030 used when an argument is passed in multiple locations. The mode of the
4031 @code{parallel} should be the mode of the entire argument. The
4032 @code{parallel} holds any number of @code{expr_list} pairs; each one
4033 describes where part of the argument is passed. In each
4034 @code{expr_list} the first operand must be a @code{reg} RTX for the hard
4035 register in which to pass this part of the argument, and the mode of the
4036 register RTX indicates how large this part of the argument is. The
4037 second operand of the @code{expr_list} is a @code{const_int} which gives
4038 the offset in bytes into the entire argument of where this part starts.
4039 As a special exception the first @code{expr_list} in the @code{parallel}
4040 RTX may have a first operand of zero. This indicates that the entire
4041 argument is also stored on the stack.
4043 The last time this macro is called, it is called with @code{MODE ==
4044 VOIDmode}, and its result is passed to the @code{call} or @code{call_value}
4045 pattern as operands 2 and 3 respectively.
4047 @cindex @file{stdarg.h} and register arguments
4048 The usual way to make the ISO library @file{stdarg.h} work on a machine
4049 where some arguments are usually passed in registers, is to cause
4050 nameless arguments to be passed on the stack instead. This is done
4051 by making @code{FUNCTION_ARG} return 0 whenever @var{named} is 0.
4053 @cindex @code{TARGET_MUST_PASS_IN_STACK}, and @code{FUNCTION_ARG}
4054 @cindex @code{REG_PARM_STACK_SPACE}, and @code{FUNCTION_ARG}
4055 You may use the hook @code{targetm.calls.must_pass_in_stack}
4056 in the definition of this macro to determine if this argument is of a
4057 type that must be passed in the stack. If @code{REG_PARM_STACK_SPACE}
4058 is not defined and @code{FUNCTION_ARG} returns nonzero for such an
4059 argument, the compiler will abort. If @code{REG_PARM_STACK_SPACE} is
4060 defined, the argument will be computed in the stack and then loaded into
4064 @deftypefn {Target Hook} bool TARGET_MUST_PASS_IN_STACK (enum machine_mode @var{mode}, const_tree @var{type})
4065 This target hook should return @code{true} if we should not pass @var{type}
4066 solely in registers. The file @file{expr.h} defines a
4067 definition that is usually appropriate, refer to @file{expr.h} for additional
4071 @defmac FUNCTION_INCOMING_ARG (@var{cum}, @var{mode}, @var{type}, @var{named})
4072 Define this macro if the target machine has ``register windows'', so
4073 that the register in which a function sees an arguments is not
4074 necessarily the same as the one in which the caller passed the
4077 For such machines, @code{FUNCTION_ARG} computes the register in which
4078 the caller passes the value, and @code{FUNCTION_INCOMING_ARG} should
4079 be defined in a similar fashion to tell the function being called
4080 where the arguments will arrive.
4082 If @code{FUNCTION_INCOMING_ARG} is not defined, @code{FUNCTION_ARG}
4083 serves both purposes.
4086 @deftypefn {Target Hook} int TARGET_ARG_PARTIAL_BYTES (CUMULATIVE_ARGS *@var{cum}, enum machine_mode @var{mode}, tree @var{type}, bool @var{named})
4087 This target hook returns the number of bytes at the beginning of an
4088 argument that must be put in registers. The value must be zero for
4089 arguments that are passed entirely in registers or that are entirely
4090 pushed on the stack.
4092 On some machines, certain arguments must be passed partially in
4093 registers and partially in memory. On these machines, typically the
4094 first few words of arguments are passed in registers, and the rest
4095 on the stack. If a multi-word argument (a @code{double} or a
4096 structure) crosses that boundary, its first few words must be passed
4097 in registers and the rest must be pushed. This macro tells the
4098 compiler when this occurs, and how many bytes should go in registers.
4100 @code{FUNCTION_ARG} for these arguments should return the first
4101 register to be used by the caller for this argument; likewise
4102 @code{FUNCTION_INCOMING_ARG}, for the called function.
4105 @deftypefn {Target Hook} bool TARGET_PASS_BY_REFERENCE (CUMULATIVE_ARGS *@var{cum}, enum machine_mode @var{mode}, const_tree @var{type}, bool @var{named})
4106 This target hook should return @code{true} if an argument at the
4107 position indicated by @var{cum} should be passed by reference. This
4108 predicate is queried after target independent reasons for being
4109 passed by reference, such as @code{TREE_ADDRESSABLE (type)}.
4111 If the hook returns true, a copy of that argument is made in memory and a
4112 pointer to the argument is passed instead of the argument itself.
4113 The pointer is passed in whatever way is appropriate for passing a pointer
4117 @deftypefn {Target Hook} bool TARGET_CALLEE_COPIES (CUMULATIVE_ARGS *@var{cum}, enum machine_mode @var{mode}, const_tree @var{type}, bool @var{named})
4118 The function argument described by the parameters to this hook is
4119 known to be passed by reference. The hook should return true if the
4120 function argument should be copied by the callee instead of copied
4123 For any argument for which the hook returns true, if it can be
4124 determined that the argument is not modified, then a copy need
4127 The default version of this hook always returns false.
4130 @defmac CUMULATIVE_ARGS
4131 A C type for declaring a variable that is used as the first argument of
4132 @code{FUNCTION_ARG} and other related values. For some target machines,
4133 the type @code{int} suffices and can hold the number of bytes of
4136 There is no need to record in @code{CUMULATIVE_ARGS} anything about the
4137 arguments that have been passed on the stack. The compiler has other
4138 variables to keep track of that. For target machines on which all
4139 arguments are passed on the stack, there is no need to store anything in
4140 @code{CUMULATIVE_ARGS}; however, the data structure must exist and
4141 should not be empty, so use @code{int}.
4144 @defmac OVERRIDE_ABI_FORMAT (@var{fndecl})
4145 If defined, this macro is called before generating any code for a
4146 function, but after the @var{cfun} descriptor for the function has been
4147 created. The back end may use this macro to update @var{cfun} to
4148 reflect an ABI other than that which would normally be used by default.
4149 If the compiler is generating code for a compiler-generated function,
4150 @var{fndecl} may be @code{NULL}.
4153 @defmac INIT_CUMULATIVE_ARGS (@var{cum}, @var{fntype}, @var{libname}, @var{fndecl}, @var{n_named_args})
4154 A C statement (sans semicolon) for initializing the variable
4155 @var{cum} for the state at the beginning of the argument list. The
4156 variable has type @code{CUMULATIVE_ARGS}. The value of @var{fntype}
4157 is the tree node for the data type of the function which will receive
4158 the args, or 0 if the args are to a compiler support library function.
4159 For direct calls that are not libcalls, @var{fndecl} contain the
4160 declaration node of the function. @var{fndecl} is also set when
4161 @code{INIT_CUMULATIVE_ARGS} is used to find arguments for the function
4162 being compiled. @var{n_named_args} is set to the number of named
4163 arguments, including a structure return address if it is passed as a
4164 parameter, when making a call. When processing incoming arguments,
4165 @var{n_named_args} is set to @minus{}1.
4167 When processing a call to a compiler support library function,
4168 @var{libname} identifies which one. It is a @code{symbol_ref} rtx which
4169 contains the name of the function, as a string. @var{libname} is 0 when
4170 an ordinary C function call is being processed. Thus, each time this
4171 macro is called, either @var{libname} or @var{fntype} is nonzero, but
4172 never both of them at once.
4175 @defmac INIT_CUMULATIVE_LIBCALL_ARGS (@var{cum}, @var{mode}, @var{libname})
4176 Like @code{INIT_CUMULATIVE_ARGS} but only used for outgoing libcalls,
4177 it gets a @code{MODE} argument instead of @var{fntype}, that would be
4178 @code{NULL}. @var{indirect} would always be zero, too. If this macro
4179 is not defined, @code{INIT_CUMULATIVE_ARGS (cum, NULL_RTX, libname,
4180 0)} is used instead.
4183 @defmac INIT_CUMULATIVE_INCOMING_ARGS (@var{cum}, @var{fntype}, @var{libname})
4184 Like @code{INIT_CUMULATIVE_ARGS} but overrides it for the purposes of
4185 finding the arguments for the function being compiled. If this macro is
4186 undefined, @code{INIT_CUMULATIVE_ARGS} is used instead.
4188 The value passed for @var{libname} is always 0, since library routines
4189 with special calling conventions are never compiled with GCC@. The
4190 argument @var{libname} exists for symmetry with
4191 @code{INIT_CUMULATIVE_ARGS}.
4192 @c could use "this macro" in place of @code{INIT_CUMULATIVE_ARGS}, maybe.
4193 @c --mew 5feb93 i switched the order of the sentences. --mew 10feb93
4196 @defmac FUNCTION_ARG_ADVANCE (@var{cum}, @var{mode}, @var{type}, @var{named})
4197 A C statement (sans semicolon) to update the summarizer variable
4198 @var{cum} to advance past an argument in the argument list. The
4199 values @var{mode}, @var{type} and @var{named} describe that argument.
4200 Once this is done, the variable @var{cum} is suitable for analyzing
4201 the @emph{following} argument with @code{FUNCTION_ARG}, etc.
4203 This macro need not do anything if the argument in question was passed
4204 on the stack. The compiler knows how to track the amount of stack space
4205 used for arguments without any special help.
4208 @defmac FUNCTION_ARG_OFFSET (@var{mode}, @var{type})
4209 If defined, a C expression that is the number of bytes to add to the
4210 offset of the argument passed in memory. This is needed for the SPU,
4211 which passes @code{char} and @code{short} arguments in the preferred
4212 slot that is in the middle of the quad word instead of starting at the
4216 @defmac FUNCTION_ARG_PADDING (@var{mode}, @var{type})
4217 If defined, a C expression which determines whether, and in which direction,
4218 to pad out an argument with extra space. The value should be of type
4219 @code{enum direction}: either @code{upward} to pad above the argument,
4220 @code{downward} to pad below, or @code{none} to inhibit padding.
4222 The @emph{amount} of padding is always just enough to reach the next
4223 multiple of @code{FUNCTION_ARG_BOUNDARY}; this macro does not control
4226 This macro has a default definition which is right for most systems.
4227 For little-endian machines, the default is to pad upward. For
4228 big-endian machines, the default is to pad downward for an argument of
4229 constant size shorter than an @code{int}, and upward otherwise.
4232 @defmac PAD_VARARGS_DOWN
4233 If defined, a C expression which determines whether the default
4234 implementation of va_arg will attempt to pad down before reading the
4235 next argument, if that argument is smaller than its aligned space as
4236 controlled by @code{PARM_BOUNDARY}. If this macro is not defined, all such
4237 arguments are padded down if @code{BYTES_BIG_ENDIAN} is true.
4240 @defmac BLOCK_REG_PADDING (@var{mode}, @var{type}, @var{first})
4241 Specify padding for the last element of a block move between registers and
4242 memory. @var{first} is nonzero if this is the only element. Defining this
4243 macro allows better control of register function parameters on big-endian
4244 machines, without using @code{PARALLEL} rtl. In particular,
4245 @code{MUST_PASS_IN_STACK} need not test padding and mode of types in
4246 registers, as there is no longer a "wrong" part of a register; For example,
4247 a three byte aggregate may be passed in the high part of a register if so
4251 @defmac FUNCTION_ARG_BOUNDARY (@var{mode}, @var{type})
4252 If defined, a C expression that gives the alignment boundary, in bits,
4253 of an argument with the specified mode and type. If it is not defined,
4254 @code{PARM_BOUNDARY} is used for all arguments.
4257 @defmac FUNCTION_ARG_REGNO_P (@var{regno})
4258 A C expression that is nonzero if @var{regno} is the number of a hard
4259 register in which function arguments are sometimes passed. This does
4260 @emph{not} include implicit arguments such as the static chain and
4261 the structure-value address. On many machines, no registers can be
4262 used for this purpose since all function arguments are pushed on the
4266 @deftypefn {Target Hook} bool TARGET_SPLIT_COMPLEX_ARG (const_tree @var{type})
4267 This hook should return true if parameter of type @var{type} are passed
4268 as two scalar parameters. By default, GCC will attempt to pack complex
4269 arguments into the target's word size. Some ABIs require complex arguments
4270 to be split and treated as their individual components. For example, on
4271 AIX64, complex floats should be passed in a pair of floating point
4272 registers, even though a complex float would fit in one 64-bit floating
4275 The default value of this hook is @code{NULL}, which is treated as always
4279 @deftypefn {Target Hook} tree TARGET_BUILD_BUILTIN_VA_LIST (void)
4280 This hook returns a type node for @code{va_list} for the target.
4281 The default version of the hook returns @code{void*}.
4284 @deftypefn {Target Hook} int TARGET_ENUM_VA_LIST_P (int @var{idx}, const char **@var{pname}, tree *@var{ptree})
4285 This target hook is used in function @code{c_common_nodes_and_builtins}
4286 to iterate through the target specific builtin types for va_list. The
4287 variable @var{idx} is used as iterator. @var{pname} has to be a pointer
4288 to a @code{const char *} and @var{ptree} a pointer to a @code{tree} typed
4290 The arguments @var{pname} and @var{ptree} are used to store the result of
4291 this macro and are set to the name of the va_list builtin type and its
4293 If the return value of this macro is zero, then there is no more element.
4294 Otherwise the @var{IDX} should be increased for the next call of this
4295 macro to iterate through all types.
4298 @deftypefn {Target Hook} tree TARGET_FN_ABI_VA_LIST (tree @var{fndecl})
4299 This hook returns the va_list type of the calling convention specified by
4301 The default version of this hook returns @code{va_list_type_node}.
4304 @deftypefn {Target Hook} tree TARGET_CANONICAL_VA_LIST_TYPE (tree @var{type})
4305 This hook returns the va_list type of the calling convention specified by the
4306 type of @var{type}. If @var{type} is not a valid va_list type, it returns
4310 @deftypefn {Target Hook} tree TARGET_GIMPLIFY_VA_ARG_EXPR (tree @var{valist}, tree @var{type}, gimple_seq *@var{pre_p}, gimple_seq *@var{post_p})
4311 This hook performs target-specific gimplification of
4312 @code{VA_ARG_EXPR}. The first two parameters correspond to the
4313 arguments to @code{va_arg}; the latter two are as in
4314 @code{gimplify.c:gimplify_expr}.
4317 @deftypefn {Target Hook} bool TARGET_VALID_POINTER_MODE (enum machine_mode @var{mode})
4318 Define this to return nonzero if the port can handle pointers
4319 with machine mode @var{mode}. The default version of this
4320 hook returns true for both @code{ptr_mode} and @code{Pmode}.
4323 @deftypefn {Target Hook} bool TARGET_SCALAR_MODE_SUPPORTED_P (enum machine_mode @var{mode})
4324 Define this to return nonzero if the port is prepared to handle
4325 insns involving scalar mode @var{mode}. For a scalar mode to be
4326 considered supported, all the basic arithmetic and comparisons
4329 The default version of this hook returns true for any mode
4330 required to handle the basic C types (as defined by the port).
4331 Included here are the double-word arithmetic supported by the
4332 code in @file{optabs.c}.
4335 @deftypefn {Target Hook} bool TARGET_VECTOR_MODE_SUPPORTED_P (enum machine_mode @var{mode})
4336 Define this to return nonzero if the port is prepared to handle
4337 insns involving vector mode @var{mode}. At the very least, it
4338 must have move patterns for this mode.
4341 @deftypefn {Target Hook} bool TARGET_SMALL_REGISTER_CLASSES_FOR_MODE_P (enum machine_mode @var{mode})
4342 Define this to return nonzero for machine modes for which the port has
4343 small register classes. If this target hook returns nonzero for a given
4344 @var{mode}, the compiler will try to minimize the lifetime of registers
4345 in @var{mode}. The hook may be called with @code{VOIDmode} as argument.
4346 In this case, the hook is expected to return nonzero if it returns nonzero
4349 On some machines, it is risky to let hard registers live across arbitrary
4350 insns. Typically, these machines have instructions that require values
4351 to be in specific registers (like an accumulator), and reload will fail
4352 if the required hard register is used for another purpose across such an
4355 Passes before reload do not know which hard registers will be used
4356 in an instruction, but the machine modes of the registers set or used in
4357 the instruction are already known. And for some machines, register
4358 classes are small for, say, integer registers but not for floating point
4359 registers. For example, the AMD x86-64 architecture requires specific
4360 registers for the legacy x86 integer instructions, but there are many
4361 SSE registers for floating point operations. On such targets, a good
4362 strategy may be to return nonzero from this hook for @code{INTEGRAL_MODE_P}
4363 machine modes but zero for the SSE register classes.
4365 The default version of this hook retuns false for any mode. It is always
4366 safe to redefine this hook to return with a nonzero value. But if you
4367 unnecessarily define it, you will reduce the amount of optimizations
4368 that can be performed in some cases. If you do not define this hook
4369 to return a nonzero value when it is required, the compiler will run out
4370 of spill registers and print a fatal error message.
4374 @subsection How Scalar Function Values Are Returned
4375 @cindex return values in registers
4376 @cindex values, returned by functions
4377 @cindex scalars, returned as values
4379 This section discusses the macros that control returning scalars as
4380 values---values that can fit in registers.
4382 @deftypefn {Target Hook} rtx TARGET_FUNCTION_VALUE (const_tree @var{ret_type}, const_tree @var{fn_decl_or_type}, bool @var{outgoing})
4384 Define this to return an RTX representing the place where a function
4385 returns or receives a value of data type @var{ret_type}, a tree node
4386 representing a data type. @var{fn_decl_or_type} is a tree node
4387 representing @code{FUNCTION_DECL} or @code{FUNCTION_TYPE} of a
4388 function being called. If @var{outgoing} is false, the hook should
4389 compute the register in which the caller will see the return value.
4390 Otherwise, the hook should return an RTX representing the place where
4391 a function returns a value.
4393 On many machines, only @code{TYPE_MODE (@var{ret_type})} is relevant.
4394 (Actually, on most machines, scalar values are returned in the same
4395 place regardless of mode.) The value of the expression is usually a
4396 @code{reg} RTX for the hard register where the return value is stored.
4397 The value can also be a @code{parallel} RTX, if the return value is in
4398 multiple places. See @code{FUNCTION_ARG} for an explanation of the
4399 @code{parallel} form. Note that the callee will populate every
4400 location specified in the @code{parallel}, but if the first element of
4401 the @code{parallel} contains the whole return value, callers will use
4402 that element as the canonical location and ignore the others. The m68k
4403 port uses this type of @code{parallel} to return pointers in both
4404 @samp{%a0} (the canonical location) and @samp{%d0}.
4406 If @code{TARGET_PROMOTE_FUNCTION_RETURN} returns true, you must apply
4407 the same promotion rules specified in @code{PROMOTE_MODE} if
4408 @var{valtype} is a scalar type.
4410 If the precise function being called is known, @var{func} is a tree
4411 node (@code{FUNCTION_DECL}) for it; otherwise, @var{func} is a null
4412 pointer. This makes it possible to use a different value-returning
4413 convention for specific functions when all their calls are
4416 Some target machines have ``register windows'' so that the register in
4417 which a function returns its value is not the same as the one in which
4418 the caller sees the value. For such machines, you should return
4419 different RTX depending on @var{outgoing}.
4421 @code{TARGET_FUNCTION_VALUE} is not used for return values with
4422 aggregate data types, because these are returned in another way. See
4423 @code{TARGET_STRUCT_VALUE_RTX} and related macros, below.
4426 @defmac FUNCTION_VALUE (@var{valtype}, @var{func})
4427 This macro has been deprecated. Use @code{TARGET_FUNCTION_VALUE} for
4428 a new target instead.
4431 @defmac LIBCALL_VALUE (@var{mode})
4432 A C expression to create an RTX representing the place where a library
4433 function returns a value of mode @var{mode}.
4435 Note that ``library function'' in this context means a compiler
4436 support routine, used to perform arithmetic, whose name is known
4437 specially by the compiler and was not mentioned in the C code being
4441 @deftypefn {Target Hook} rtx TARGET_LIBCALL_VALUE (enum machine_mode @var{mode}, const_rtx @var{fun})
4442 Define this hook if the back-end needs to know the name of the libcall
4443 function in order to determine where the result should be returned.
4445 The mode of the result is given by @var{mode} and the name of the called
4446 library function is given by @var{fun}. The hook should return an RTX
4447 representing the place where the library function result will be returned.
4449 If this hook is not defined, then LIBCALL_VALUE will be used.
4452 @defmac FUNCTION_VALUE_REGNO_P (@var{regno})
4453 A C expression that is nonzero if @var{regno} is the number of a hard
4454 register in which the values of called function may come back.
4456 A register whose use for returning values is limited to serving as the
4457 second of a pair (for a value of type @code{double}, say) need not be
4458 recognized by this macro. So for most machines, this definition
4462 #define FUNCTION_VALUE_REGNO_P(N) ((N) == 0)
4465 If the machine has register windows, so that the caller and the called
4466 function use different registers for the return value, this macro
4467 should recognize only the caller's register numbers.
4469 This macro has been deprecated. Use @code{TARGET_FUNCTION_VALUE_REGNO_P}
4470 for a new target instead.
4473 @deftypefn {Target Hook} bool TARGET_FUNCTION_VALUE_REGNO_P (const unsigned int @var{regno})
4474 A target hook that return @code{true} if @var{regno} is the number of a hard
4475 register in which the values of called function may come back.
4477 A register whose use for returning values is limited to serving as the
4478 second of a pair (for a value of type @code{double}, say) need not be
4479 recognized by this target hook.
4481 If the machine has register windows, so that the caller and the called
4482 function use different registers for the return value, this target hook
4483 should recognize only the caller's register numbers.
4485 If this hook is not defined, then FUNCTION_VALUE_REGNO_P will be used.
4488 @defmac APPLY_RESULT_SIZE
4489 Define this macro if @samp{untyped_call} and @samp{untyped_return}
4490 need more space than is implied by @code{FUNCTION_VALUE_REGNO_P} for
4491 saving and restoring an arbitrary return value.
4494 @deftypefn {Target Hook} bool TARGET_RETURN_IN_MSB (const_tree @var{type})
4495 This hook should return true if values of type @var{type} are returned
4496 at the most significant end of a register (in other words, if they are
4497 padded at the least significant end). You can assume that @var{type}
4498 is returned in a register; the caller is required to check this.
4500 Note that the register provided by @code{TARGET_FUNCTION_VALUE} must
4501 be able to hold the complete return value. For example, if a 1-, 2-
4502 or 3-byte structure is returned at the most significant end of a
4503 4-byte register, @code{TARGET_FUNCTION_VALUE} should provide an
4507 @node Aggregate Return
4508 @subsection How Large Values Are Returned
4509 @cindex aggregates as return values
4510 @cindex large return values
4511 @cindex returning aggregate values
4512 @cindex structure value address
4514 When a function value's mode is @code{BLKmode} (and in some other
4515 cases), the value is not returned according to
4516 @code{TARGET_FUNCTION_VALUE} (@pxref{Scalar Return}). Instead, the
4517 caller passes the address of a block of memory in which the value
4518 should be stored. This address is called the @dfn{structure value
4521 This section describes how to control returning structure values in
4524 @deftypefn {Target Hook} bool TARGET_RETURN_IN_MEMORY (const_tree @var{type}, const_tree @var{fntype})
4525 This target hook should return a nonzero value to say to return the
4526 function value in memory, just as large structures are always returned.
4527 Here @var{type} will be the data type of the value, and @var{fntype}
4528 will be the type of the function doing the returning, or @code{NULL} for
4531 Note that values of mode @code{BLKmode} must be explicitly handled
4532 by this function. Also, the option @option{-fpcc-struct-return}
4533 takes effect regardless of this macro. On most systems, it is
4534 possible to leave the hook undefined; this causes a default
4535 definition to be used, whose value is the constant 1 for @code{BLKmode}
4536 values, and 0 otherwise.
4538 Do not use this hook to indicate that structures and unions should always
4539 be returned in memory. You should instead use @code{DEFAULT_PCC_STRUCT_RETURN}
4543 @defmac DEFAULT_PCC_STRUCT_RETURN
4544 Define this macro to be 1 if all structure and union return values must be
4545 in memory. Since this results in slower code, this should be defined
4546 only if needed for compatibility with other compilers or with an ABI@.
4547 If you define this macro to be 0, then the conventions used for structure
4548 and union return values are decided by the @code{TARGET_RETURN_IN_MEMORY}
4551 If not defined, this defaults to the value 1.
4554 @deftypefn {Target Hook} rtx TARGET_STRUCT_VALUE_RTX (tree @var{fndecl}, int @var{incoming})
4555 This target hook should return the location of the structure value
4556 address (normally a @code{mem} or @code{reg}), or 0 if the address is
4557 passed as an ``invisible'' first argument. Note that @var{fndecl} may
4558 be @code{NULL}, for libcalls. You do not need to define this target
4559 hook if the address is always passed as an ``invisible'' first
4562 On some architectures the place where the structure value address
4563 is found by the called function is not the same place that the
4564 caller put it. This can be due to register windows, or it could
4565 be because the function prologue moves it to a different place.
4566 @var{incoming} is @code{1} or @code{2} when the location is needed in
4567 the context of the called function, and @code{0} in the context of
4570 If @var{incoming} is nonzero and the address is to be found on the
4571 stack, return a @code{mem} which refers to the frame pointer. If
4572 @var{incoming} is @code{2}, the result is being used to fetch the
4573 structure value address at the beginning of a function. If you need
4574 to emit adjusting code, you should do it at this point.
4577 @defmac PCC_STATIC_STRUCT_RETURN
4578 Define this macro if the usual system convention on the target machine
4579 for returning structures and unions is for the called function to return
4580 the address of a static variable containing the value.
4582 Do not define this if the usual system convention is for the caller to
4583 pass an address to the subroutine.
4585 This macro has effect in @option{-fpcc-struct-return} mode, but it does
4586 nothing when you use @option{-freg-struct-return} mode.
4590 @subsection Caller-Saves Register Allocation
4592 If you enable it, GCC can save registers around function calls. This
4593 makes it possible to use call-clobbered registers to hold variables that
4594 must live across calls.
4596 @defmac CALLER_SAVE_PROFITABLE (@var{refs}, @var{calls})
4597 A C expression to determine whether it is worthwhile to consider placing
4598 a pseudo-register in a call-clobbered hard register and saving and
4599 restoring it around each function call. The expression should be 1 when
4600 this is worth doing, and 0 otherwise.
4602 If you don't define this macro, a default is used which is good on most
4603 machines: @code{4 * @var{calls} < @var{refs}}.
4606 @defmac HARD_REGNO_CALLER_SAVE_MODE (@var{regno}, @var{nregs})
4607 A C expression specifying which mode is required for saving @var{nregs}
4608 of a pseudo-register in call-clobbered hard register @var{regno}. If
4609 @var{regno} is unsuitable for caller save, @code{VOIDmode} should be
4610 returned. For most machines this macro need not be defined since GCC
4611 will select the smallest suitable mode.
4614 @node Function Entry
4615 @subsection Function Entry and Exit
4616 @cindex function entry and exit
4620 This section describes the macros that output function entry
4621 (@dfn{prologue}) and exit (@dfn{epilogue}) code.
4623 @deftypefn {Target Hook} void TARGET_ASM_FUNCTION_PROLOGUE (FILE *@var{file}, HOST_WIDE_INT @var{size})
4624 If defined, a function that outputs the assembler code for entry to a
4625 function. The prologue is responsible for setting up the stack frame,
4626 initializing the frame pointer register, saving registers that must be
4627 saved, and allocating @var{size} additional bytes of storage for the
4628 local variables. @var{size} is an integer. @var{file} is a stdio
4629 stream to which the assembler code should be output.
4631 The label for the beginning of the function need not be output by this
4632 macro. That has already been done when the macro is run.
4634 @findex regs_ever_live
4635 To determine which registers to save, the macro can refer to the array
4636 @code{regs_ever_live}: element @var{r} is nonzero if hard register
4637 @var{r} is used anywhere within the function. This implies the function
4638 prologue should save register @var{r}, provided it is not one of the
4639 call-used registers. (@code{TARGET_ASM_FUNCTION_EPILOGUE} must likewise use
4640 @code{regs_ever_live}.)
4642 On machines that have ``register windows'', the function entry code does
4643 not save on the stack the registers that are in the windows, even if
4644 they are supposed to be preserved by function calls; instead it takes
4645 appropriate steps to ``push'' the register stack, if any non-call-used
4646 registers are used in the function.
4648 @findex frame_pointer_needed
4649 On machines where functions may or may not have frame-pointers, the
4650 function entry code must vary accordingly; it must set up the frame
4651 pointer if one is wanted, and not otherwise. To determine whether a
4652 frame pointer is in wanted, the macro can refer to the variable
4653 @code{frame_pointer_needed}. The variable's value will be 1 at run
4654 time in a function that needs a frame pointer. @xref{Elimination}.
4656 The function entry code is responsible for allocating any stack space
4657 required for the function. This stack space consists of the regions
4658 listed below. In most cases, these regions are allocated in the
4659 order listed, with the last listed region closest to the top of the
4660 stack (the lowest address if @code{STACK_GROWS_DOWNWARD} is defined, and
4661 the highest address if it is not defined). You can use a different order
4662 for a machine if doing so is more convenient or required for
4663 compatibility reasons. Except in cases where required by standard
4664 or by a debugger, there is no reason why the stack layout used by GCC
4665 need agree with that used by other compilers for a machine.
4668 @deftypefn {Target Hook} void TARGET_ASM_FUNCTION_END_PROLOGUE (FILE *@var{file})
4669 If defined, a function that outputs assembler code at the end of a
4670 prologue. This should be used when the function prologue is being
4671 emitted as RTL, and you have some extra assembler that needs to be
4672 emitted. @xref{prologue instruction pattern}.
4675 @deftypefn {Target Hook} void TARGET_ASM_FUNCTION_BEGIN_EPILOGUE (FILE *@var{file})
4676 If defined, a function that outputs assembler code at the start of an
4677 epilogue. This should be used when the function epilogue is being
4678 emitted as RTL, and you have some extra assembler that needs to be
4679 emitted. @xref{epilogue instruction pattern}.
4682 @deftypefn {Target Hook} void TARGET_ASM_FUNCTION_EPILOGUE (FILE *@var{file}, HOST_WIDE_INT @var{size})
4683 If defined, a function that outputs the assembler code for exit from a
4684 function. The epilogue is responsible for restoring the saved
4685 registers and stack pointer to their values when the function was
4686 called, and returning control to the caller. This macro takes the
4687 same arguments as the macro @code{TARGET_ASM_FUNCTION_PROLOGUE}, and the
4688 registers to restore are determined from @code{regs_ever_live} and
4689 @code{CALL_USED_REGISTERS} in the same way.
4691 On some machines, there is a single instruction that does all the work
4692 of returning from the function. On these machines, give that
4693 instruction the name @samp{return} and do not define the macro
4694 @code{TARGET_ASM_FUNCTION_EPILOGUE} at all.
4696 Do not define a pattern named @samp{return} if you want the
4697 @code{TARGET_ASM_FUNCTION_EPILOGUE} to be used. If you want the target
4698 switches to control whether return instructions or epilogues are used,
4699 define a @samp{return} pattern with a validity condition that tests the
4700 target switches appropriately. If the @samp{return} pattern's validity
4701 condition is false, epilogues will be used.
4703 On machines where functions may or may not have frame-pointers, the
4704 function exit code must vary accordingly. Sometimes the code for these
4705 two cases is completely different. To determine whether a frame pointer
4706 is wanted, the macro can refer to the variable
4707 @code{frame_pointer_needed}. The variable's value will be 1 when compiling
4708 a function that needs a frame pointer.
4710 Normally, @code{TARGET_ASM_FUNCTION_PROLOGUE} and
4711 @code{TARGET_ASM_FUNCTION_EPILOGUE} must treat leaf functions specially.
4712 The C variable @code{current_function_is_leaf} is nonzero for such a
4713 function. @xref{Leaf Functions}.
4715 On some machines, some functions pop their arguments on exit while
4716 others leave that for the caller to do. For example, the 68020 when
4717 given @option{-mrtd} pops arguments in functions that take a fixed
4718 number of arguments.
4720 @findex current_function_pops_args
4721 Your definition of the macro @code{RETURN_POPS_ARGS} decides which
4722 functions pop their own arguments. @code{TARGET_ASM_FUNCTION_EPILOGUE}
4723 needs to know what was decided. The number of bytes of the current
4724 function's arguments that this function should pop is available in
4725 @code{crtl->args.pops_args}. @xref{Scalar Return}.
4730 @findex current_function_pretend_args_size
4731 A region of @code{current_function_pretend_args_size} bytes of
4732 uninitialized space just underneath the first argument arriving on the
4733 stack. (This may not be at the very start of the allocated stack region
4734 if the calling sequence has pushed anything else since pushing the stack
4735 arguments. But usually, on such machines, nothing else has been pushed
4736 yet, because the function prologue itself does all the pushing.) This
4737 region is used on machines where an argument may be passed partly in
4738 registers and partly in memory, and, in some cases to support the
4739 features in @code{<stdarg.h>}.
4742 An area of memory used to save certain registers used by the function.
4743 The size of this area, which may also include space for such things as
4744 the return address and pointers to previous stack frames, is
4745 machine-specific and usually depends on which registers have been used
4746 in the function. Machines with register windows often do not require
4750 A region of at least @var{size} bytes, possibly rounded up to an allocation
4751 boundary, to contain the local variables of the function. On some machines,
4752 this region and the save area may occur in the opposite order, with the
4753 save area closer to the top of the stack.
4756 @cindex @code{ACCUMULATE_OUTGOING_ARGS} and stack frames
4757 Optionally, when @code{ACCUMULATE_OUTGOING_ARGS} is defined, a region of
4758 @code{current_function_outgoing_args_size} bytes to be used for outgoing
4759 argument lists of the function. @xref{Stack Arguments}.
4762 @defmac EXIT_IGNORE_STACK
4763 Define this macro as a C expression that is nonzero if the return
4764 instruction or the function epilogue ignores the value of the stack
4765 pointer; in other words, if it is safe to delete an instruction to
4766 adjust the stack pointer before a return from the function. The
4769 Note that this macro's value is relevant only for functions for which
4770 frame pointers are maintained. It is never safe to delete a final
4771 stack adjustment in a function that has no frame pointer, and the
4772 compiler knows this regardless of @code{EXIT_IGNORE_STACK}.
4775 @defmac EPILOGUE_USES (@var{regno})
4776 Define this macro as a C expression that is nonzero for registers that are
4777 used by the epilogue or the @samp{return} pattern. The stack and frame
4778 pointer registers are already assumed to be used as needed.
4781 @defmac EH_USES (@var{regno})
4782 Define this macro as a C expression that is nonzero for registers that are
4783 used by the exception handling mechanism, and so should be considered live
4784 on entry to an exception edge.
4787 @defmac DELAY_SLOTS_FOR_EPILOGUE
4788 Define this macro if the function epilogue contains delay slots to which
4789 instructions from the rest of the function can be ``moved''. The
4790 definition should be a C expression whose value is an integer
4791 representing the number of delay slots there.
4794 @defmac ELIGIBLE_FOR_EPILOGUE_DELAY (@var{insn}, @var{n})
4795 A C expression that returns 1 if @var{insn} can be placed in delay
4796 slot number @var{n} of the epilogue.
4798 The argument @var{n} is an integer which identifies the delay slot now
4799 being considered (since different slots may have different rules of
4800 eligibility). It is never negative and is always less than the number
4801 of epilogue delay slots (what @code{DELAY_SLOTS_FOR_EPILOGUE} returns).
4802 If you reject a particular insn for a given delay slot, in principle, it
4803 may be reconsidered for a subsequent delay slot. Also, other insns may
4804 (at least in principle) be considered for the so far unfilled delay
4807 @findex current_function_epilogue_delay_list
4808 @findex final_scan_insn
4809 The insns accepted to fill the epilogue delay slots are put in an RTL
4810 list made with @code{insn_list} objects, stored in the variable
4811 @code{current_function_epilogue_delay_list}. The insn for the first
4812 delay slot comes first in the list. Your definition of the macro
4813 @code{TARGET_ASM_FUNCTION_EPILOGUE} should fill the delay slots by
4814 outputting the insns in this list, usually by calling
4815 @code{final_scan_insn}.
4817 You need not define this macro if you did not define
4818 @code{DELAY_SLOTS_FOR_EPILOGUE}.
4821 @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})
4822 A function that outputs the assembler code for a thunk
4823 function, used to implement C++ virtual function calls with multiple
4824 inheritance. The thunk acts as a wrapper around a virtual function,
4825 adjusting the implicit object parameter before handing control off to
4828 First, emit code to add the integer @var{delta} to the location that
4829 contains the incoming first argument. Assume that this argument
4830 contains a pointer, and is the one used to pass the @code{this} pointer
4831 in C++. This is the incoming argument @emph{before} the function prologue,
4832 e.g.@: @samp{%o0} on a sparc. The addition must preserve the values of
4833 all other incoming arguments.
4835 Then, if @var{vcall_offset} is nonzero, an additional adjustment should be
4836 made after adding @code{delta}. In particular, if @var{p} is the
4837 adjusted pointer, the following adjustment should be made:
4840 p += (*((ptrdiff_t **)p))[vcall_offset/sizeof(ptrdiff_t)]
4843 After the additions, emit code to jump to @var{function}, which is a
4844 @code{FUNCTION_DECL}. This is a direct pure jump, not a call, and does
4845 not touch the return address. Hence returning from @var{FUNCTION} will
4846 return to whoever called the current @samp{thunk}.
4848 The effect must be as if @var{function} had been called directly with
4849 the adjusted first argument. This macro is responsible for emitting all
4850 of the code for a thunk function; @code{TARGET_ASM_FUNCTION_PROLOGUE}
4851 and @code{TARGET_ASM_FUNCTION_EPILOGUE} are not invoked.
4853 The @var{thunk_fndecl} is redundant. (@var{delta} and @var{function}
4854 have already been extracted from it.) It might possibly be useful on
4855 some targets, but probably not.
4857 If you do not define this macro, the target-independent code in the C++
4858 front end will generate a less efficient heavyweight thunk that calls
4859 @var{function} instead of jumping to it. The generic approach does
4860 not support varargs.
4863 @deftypefn {Target Hook} bool TARGET_ASM_CAN_OUTPUT_MI_THUNK (const_tree @var{thunk_fndecl}, HOST_WIDE_INT @var{delta}, HOST_WIDE_INT @var{vcall_offset}, const_tree @var{function})
4864 A function that returns true if TARGET_ASM_OUTPUT_MI_THUNK would be able
4865 to output the assembler code for the thunk function specified by the
4866 arguments it is passed, and false otherwise. In the latter case, the
4867 generic approach will be used by the C++ front end, with the limitations
4872 @subsection Generating Code for Profiling
4873 @cindex profiling, code generation
4875 These macros will help you generate code for profiling.
4877 @defmac FUNCTION_PROFILER (@var{file}, @var{labelno})
4878 A C statement or compound statement to output to @var{file} some
4879 assembler code to call the profiling subroutine @code{mcount}.
4882 The details of how @code{mcount} expects to be called are determined by
4883 your operating system environment, not by GCC@. To figure them out,
4884 compile a small program for profiling using the system's installed C
4885 compiler and look at the assembler code that results.
4887 Older implementations of @code{mcount} expect the address of a counter
4888 variable to be loaded into some register. The name of this variable is
4889 @samp{LP} followed by the number @var{labelno}, so you would generate
4890 the name using @samp{LP%d} in a @code{fprintf}.
4893 @defmac PROFILE_HOOK
4894 A C statement or compound statement to output to @var{file} some assembly
4895 code to call the profiling subroutine @code{mcount} even the target does
4896 not support profiling.
4899 @defmac NO_PROFILE_COUNTERS
4900 Define this macro to be an expression with a nonzero value if the
4901 @code{mcount} subroutine on your system does not need a counter variable
4902 allocated for each function. This is true for almost all modern
4903 implementations. If you define this macro, you must not use the
4904 @var{labelno} argument to @code{FUNCTION_PROFILER}.
4907 @defmac PROFILE_BEFORE_PROLOGUE
4908 Define this macro if the code for function profiling should come before
4909 the function prologue. Normally, the profiling code comes after.
4913 @subsection Permitting tail calls
4916 @deftypefn {Target Hook} bool TARGET_FUNCTION_OK_FOR_SIBCALL (tree @var{decl}, tree @var{exp})
4917 True if it is ok to do sibling call optimization for the specified
4918 call expression @var{exp}. @var{decl} will be the called function,
4919 or @code{NULL} if this is an indirect call.
4921 It is not uncommon for limitations of calling conventions to prevent
4922 tail calls to functions outside the current unit of translation, or
4923 during PIC compilation. The hook is used to enforce these restrictions,
4924 as the @code{sibcall} md pattern can not fail, or fall over to a
4925 ``normal'' call. The criteria for successful sibling call optimization
4926 may vary greatly between different architectures.
4929 @deftypefn {Target Hook} void TARGET_EXTRA_LIVE_ON_ENTRY (bitmap @var{regs})
4930 Add any hard registers to @var{regs} that are live on entry to the
4931 function. This hook only needs to be defined to provide registers that
4932 cannot be found by examination of FUNCTION_ARG_REGNO_P, the callee saved
4933 registers, STATIC_CHAIN_INCOMING_REGNUM, STATIC_CHAIN_REGNUM,
4934 TARGET_STRUCT_VALUE_RTX, FRAME_POINTER_REGNUM, EH_USES,
4935 FRAME_POINTER_REGNUM, ARG_POINTER_REGNUM, and the PIC_OFFSET_TABLE_REGNUM.
4938 @node Stack Smashing Protection
4939 @subsection Stack smashing protection
4940 @cindex stack smashing protection
4942 @deftypefn {Target Hook} tree TARGET_STACK_PROTECT_GUARD (void)
4943 This hook returns a @code{DECL} node for the external variable to use
4944 for the stack protection guard. This variable is initialized by the
4945 runtime to some random value and is used to initialize the guard value
4946 that is placed at the top of the local stack frame. The type of this
4947 variable must be @code{ptr_type_node}.
4949 The default version of this hook creates a variable called
4950 @samp{__stack_chk_guard}, which is normally defined in @file{libgcc2.c}.
4953 @deftypefn {Target Hook} tree TARGET_STACK_PROTECT_FAIL (void)
4954 This hook returns a tree expression that alerts the runtime that the
4955 stack protect guard variable has been modified. This expression should
4956 involve a call to a @code{noreturn} function.
4958 The default version of this hook invokes a function called
4959 @samp{__stack_chk_fail}, taking no arguments. This function is
4960 normally defined in @file{libgcc2.c}.
4964 @section Implementing the Varargs Macros
4965 @cindex varargs implementation
4967 GCC comes with an implementation of @code{<varargs.h>} and
4968 @code{<stdarg.h>} that work without change on machines that pass arguments
4969 on the stack. Other machines require their own implementations of
4970 varargs, and the two machine independent header files must have
4971 conditionals to include it.
4973 ISO @code{<stdarg.h>} differs from traditional @code{<varargs.h>} mainly in
4974 the calling convention for @code{va_start}. The traditional
4975 implementation takes just one argument, which is the variable in which
4976 to store the argument pointer. The ISO implementation of
4977 @code{va_start} takes an additional second argument. The user is
4978 supposed to write the last named argument of the function here.
4980 However, @code{va_start} should not use this argument. The way to find
4981 the end of the named arguments is with the built-in functions described
4984 @defmac __builtin_saveregs ()
4985 Use this built-in function to save the argument registers in memory so
4986 that the varargs mechanism can access them. Both ISO and traditional
4987 versions of @code{va_start} must use @code{__builtin_saveregs}, unless
4988 you use @code{TARGET_SETUP_INCOMING_VARARGS} (see below) instead.
4990 On some machines, @code{__builtin_saveregs} is open-coded under the
4991 control of the target hook @code{TARGET_EXPAND_BUILTIN_SAVEREGS}. On
4992 other machines, it calls a routine written in assembler language,
4993 found in @file{libgcc2.c}.
4995 Code generated for the call to @code{__builtin_saveregs} appears at the
4996 beginning of the function, as opposed to where the call to
4997 @code{__builtin_saveregs} is written, regardless of what the code is.
4998 This is because the registers must be saved before the function starts
4999 to use them for its own purposes.
5000 @c i rewrote the first sentence above to fix an overfull hbox. --mew
5004 @defmac __builtin_next_arg (@var{lastarg})
5005 This builtin returns the address of the first anonymous stack
5006 argument, as type @code{void *}. If @code{ARGS_GROW_DOWNWARD}, it
5007 returns the address of the location above the first anonymous stack
5008 argument. Use it in @code{va_start} to initialize the pointer for
5009 fetching arguments from the stack. Also use it in @code{va_start} to
5010 verify that the second parameter @var{lastarg} is the last named argument
5011 of the current function.
5014 @defmac __builtin_classify_type (@var{object})
5015 Since each machine has its own conventions for which data types are
5016 passed in which kind of register, your implementation of @code{va_arg}
5017 has to embody these conventions. The easiest way to categorize the
5018 specified data type is to use @code{__builtin_classify_type} together
5019 with @code{sizeof} and @code{__alignof__}.
5021 @code{__builtin_classify_type} ignores the value of @var{object},
5022 considering only its data type. It returns an integer describing what
5023 kind of type that is---integer, floating, pointer, structure, and so on.
5025 The file @file{typeclass.h} defines an enumeration that you can use to
5026 interpret the values of @code{__builtin_classify_type}.
5029 These machine description macros help implement varargs:
5031 @deftypefn {Target Hook} rtx TARGET_EXPAND_BUILTIN_SAVEREGS (void)
5032 If defined, this hook produces the machine-specific code for a call to
5033 @code{__builtin_saveregs}. This code will be moved to the very
5034 beginning of the function, before any parameter access are made. The
5035 return value of this function should be an RTX that contains the value
5036 to use as the return of @code{__builtin_saveregs}.
5039 @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})
5040 This target hook offers an alternative to using
5041 @code{__builtin_saveregs} and defining the hook
5042 @code{TARGET_EXPAND_BUILTIN_SAVEREGS}. Use it to store the anonymous
5043 register arguments into the stack so that all the arguments appear to
5044 have been passed consecutively on the stack. Once this is done, you can
5045 use the standard implementation of varargs that works for machines that
5046 pass all their arguments on the stack.
5048 The argument @var{args_so_far} points to the @code{CUMULATIVE_ARGS} data
5049 structure, containing the values that are obtained after processing the
5050 named arguments. The arguments @var{mode} and @var{type} describe the
5051 last named argument---its machine mode and its data type as a tree node.
5053 The target hook should do two things: first, push onto the stack all the
5054 argument registers @emph{not} used for the named arguments, and second,
5055 store the size of the data thus pushed into the @code{int}-valued
5056 variable pointed to by @var{pretend_args_size}. The value that you
5057 store here will serve as additional offset for setting up the stack
5060 Because you must generate code to push the anonymous arguments at
5061 compile time without knowing their data types,
5062 @code{TARGET_SETUP_INCOMING_VARARGS} is only useful on machines that
5063 have just a single category of argument register and use it uniformly
5066 If the argument @var{second_time} is nonzero, it means that the
5067 arguments of the function are being analyzed for the second time. This
5068 happens for an inline function, which is not actually compiled until the
5069 end of the source file. The hook @code{TARGET_SETUP_INCOMING_VARARGS} should
5070 not generate any instructions in this case.
5073 @deftypefn {Target Hook} bool TARGET_STRICT_ARGUMENT_NAMING (CUMULATIVE_ARGS *@var{ca})
5074 Define this hook to return @code{true} if the location where a function
5075 argument is passed depends on whether or not it is a named argument.
5077 This hook controls how the @var{named} argument to @code{FUNCTION_ARG}
5078 is set for varargs and stdarg functions. If this hook returns
5079 @code{true}, the @var{named} argument is always true for named
5080 arguments, and false for unnamed arguments. If it returns @code{false},
5081 but @code{TARGET_PRETEND_OUTGOING_VARARGS_NAMED} returns @code{true},
5082 then all arguments are treated as named. Otherwise, all named arguments
5083 except the last are treated as named.
5085 You need not define this hook if it always returns @code{false}.
5088 @deftypefn {Target Hook} bool TARGET_PRETEND_OUTGOING_VARARGS_NAMED (CUMULATIVE_ARGS *@var{ca})
5089 If you need to conditionally change ABIs so that one works with
5090 @code{TARGET_SETUP_INCOMING_VARARGS}, but the other works like neither
5091 @code{TARGET_SETUP_INCOMING_VARARGS} nor @code{TARGET_STRICT_ARGUMENT_NAMING} was
5092 defined, then define this hook to return @code{true} if
5093 @code{TARGET_SETUP_INCOMING_VARARGS} is used, @code{false} otherwise.
5094 Otherwise, you should not define this hook.
5098 @section Trampolines for Nested Functions
5099 @cindex trampolines for nested functions
5100 @cindex nested functions, trampolines for
5102 A @dfn{trampoline} is a small piece of code that is created at run time
5103 when the address of a nested function is taken. It normally resides on
5104 the stack, in the stack frame of the containing function. These macros
5105 tell GCC how to generate code to allocate and initialize a
5108 The instructions in the trampoline must do two things: load a constant
5109 address into the static chain register, and jump to the real address of
5110 the nested function. On CISC machines such as the m68k, this requires
5111 two instructions, a move immediate and a jump. Then the two addresses
5112 exist in the trampoline as word-long immediate operands. On RISC
5113 machines, it is often necessary to load each address into a register in
5114 two parts. Then pieces of each address form separate immediate
5117 The code generated to initialize the trampoline must store the variable
5118 parts---the static chain value and the function address---into the
5119 immediate operands of the instructions. On a CISC machine, this is
5120 simply a matter of copying each address to a memory reference at the
5121 proper offset from the start of the trampoline. On a RISC machine, it
5122 may be necessary to take out pieces of the address and store them
5125 @deftypefn {Target Hook} void TARGET_ASM_TRAMPOLINE_TEMPLATE (FILE *@var{f})
5126 This hook is called by @code{assemble_trampoline_template} to output,
5127 on the stream @var{f}, assembler code for a block of data that contains
5128 the constant parts of a trampoline. This code should not include a
5129 label---the label is taken care of automatically.
5131 If you do not define this hook, it means no template is needed
5132 for the target. Do not define this hook on systems where the block move
5133 code to copy the trampoline into place would be larger than the code
5134 to generate it on the spot.
5137 @defmac TRAMPOLINE_SECTION
5138 Return the section into which the trampoline template is to be placed
5139 (@pxref{Sections}). The default value is @code{readonly_data_section}.
5142 @defmac TRAMPOLINE_SIZE
5143 A C expression for the size in bytes of the trampoline, as an integer.
5146 @defmac TRAMPOLINE_ALIGNMENT
5147 Alignment required for trampolines, in bits.
5149 If you don't define this macro, the value of @code{FUNCTION_ALIGNMENT}
5150 is used for aligning trampolines.
5153 @deftypefn {Target Hook} void TARGET_TRAMPOLINE_INIT (rtx @var{m_tramp}, tree @var{fndecl}, rtx @var{static_chain})
5154 This hook is called to initialize a trampoline.
5155 @var{m_tramp} is an RTX for the memory block for the trampoline; @var{fndecl}
5156 is the @code{FUNCTION_DECL} for the nested function; @var{static_chain} is an
5157 RTX for the static chain value that should be passed to the function
5160 If the target defines @code{TARGET_ASM_TRAMPOLINE_TEMPLATE}, then the
5161 first thing this hook should do is emit a block move into @var{m_tramp}
5162 from the memory block returned by @code{assemble_trampoline_template}.
5163 Note that the block move need only cover the constant parts of the
5164 trampoline. If the target isolates the variable parts of the trampoline
5165 to the end, not all @code{TRAMPOLINE_SIZE} bytes need be copied.
5167 If the target requires any other actions, such as flushing caches or
5168 enabling stack execution, these actions should be performed after
5169 initializing the trampoline proper.
5172 @deftypefn {Target Hook} rtx TARGET_TRAMPOLINE_ADJUST_ADDRESS (rtx @var{addr})
5173 This hook should perform any machine-specific adjustment in
5174 the address of the trampoline. Its argument contains the address of the
5175 memory block that was passed to @code{TARGET_TRAMPOLINE_INIT}. In case
5176 the address to be used for a function call should be different from the
5177 address at which the template was stored, the different address should
5178 be returned; otherwise @var{addr} should be returned unchanged.
5179 If this hook is not defined, @var{addr} will be used for function calls.
5182 Implementing trampolines is difficult on many machines because they have
5183 separate instruction and data caches. Writing into a stack location
5184 fails to clear the memory in the instruction cache, so when the program
5185 jumps to that location, it executes the old contents.
5187 Here are two possible solutions. One is to clear the relevant parts of
5188 the instruction cache whenever a trampoline is set up. The other is to
5189 make all trampolines identical, by having them jump to a standard
5190 subroutine. The former technique makes trampoline execution faster; the
5191 latter makes initialization faster.
5193 To clear the instruction cache when a trampoline is initialized, define
5194 the following macro.
5196 @defmac CLEAR_INSN_CACHE (@var{beg}, @var{end})
5197 If defined, expands to a C expression clearing the @emph{instruction
5198 cache} in the specified interval. The definition of this macro would
5199 typically be a series of @code{asm} statements. Both @var{beg} and
5200 @var{end} are both pointer expressions.
5203 The operating system may also require the stack to be made executable
5204 before calling the trampoline. To implement this requirement, define
5205 the following macro.
5207 @defmac ENABLE_EXECUTE_STACK
5208 Define this macro if certain operations must be performed before executing
5209 code located on the stack. The macro should expand to a series of C
5210 file-scope constructs (e.g.@: functions) and provide a unique entry point
5211 named @code{__enable_execute_stack}. The target is responsible for
5212 emitting calls to the entry point in the code, for example from the
5213 @code{TARGET_TRAMPOLINE_INIT} hook.
5216 To use a standard subroutine, define the following macro. In addition,
5217 you must make sure that the instructions in a trampoline fill an entire
5218 cache line with identical instructions, or else ensure that the
5219 beginning of the trampoline code is always aligned at the same point in
5220 its cache line. Look in @file{m68k.h} as a guide.
5222 @defmac TRANSFER_FROM_TRAMPOLINE
5223 Define this macro if trampolines need a special subroutine to do their
5224 work. The macro should expand to a series of @code{asm} statements
5225 which will be compiled with GCC@. They go in a library function named
5226 @code{__transfer_from_trampoline}.
5228 If you need to avoid executing the ordinary prologue code of a compiled
5229 C function when you jump to the subroutine, you can do so by placing a
5230 special label of your own in the assembler code. Use one @code{asm}
5231 statement to generate an assembler label, and another to make the label
5232 global. Then trampolines can use that label to jump directly to your
5233 special assembler code.
5237 @section Implicit Calls to Library Routines
5238 @cindex library subroutine names
5239 @cindex @file{libgcc.a}
5241 @c prevent bad page break with this line
5242 Here is an explanation of implicit calls to library routines.
5244 @defmac DECLARE_LIBRARY_RENAMES
5245 This macro, if defined, should expand to a piece of C code that will get
5246 expanded when compiling functions for libgcc.a. It can be used to
5247 provide alternate names for GCC's internal library functions if there
5248 are ABI-mandated names that the compiler should provide.
5251 @findex set_optab_libfunc
5252 @findex init_one_libfunc
5253 @deftypefn {Target Hook} void TARGET_INIT_LIBFUNCS (void)
5254 This hook should declare additional library routines or rename
5255 existing ones, using the functions @code{set_optab_libfunc} and
5256 @code{init_one_libfunc} defined in @file{optabs.c}.
5257 @code{init_optabs} calls this macro after initializing all the normal
5260 The default is to do nothing. Most ports don't need to define this hook.
5263 @defmac FLOAT_LIB_COMPARE_RETURNS_BOOL (@var{mode}, @var{comparison})
5264 This macro should return @code{true} if the library routine that
5265 implements the floating point comparison operator @var{comparison} in
5266 mode @var{mode} will return a boolean, and @var{false} if it will
5269 GCC's own floating point libraries return tristates from the
5270 comparison operators, so the default returns false always. Most ports
5271 don't need to define this macro.
5274 @defmac TARGET_LIB_INT_CMP_BIASED
5275 This macro should evaluate to @code{true} if the integer comparison
5276 functions (like @code{__cmpdi2}) return 0 to indicate that the first
5277 operand is smaller than the second, 1 to indicate that they are equal,
5278 and 2 to indicate that the first operand is greater than the second.
5279 If this macro evaluates to @code{false} the comparison functions return
5280 @minus{}1, 0, and 1 instead of 0, 1, and 2. If the target uses the routines
5281 in @file{libgcc.a}, you do not need to define this macro.
5284 @cindex US Software GOFAST, floating point emulation library
5285 @cindex floating point emulation library, US Software GOFAST
5286 @cindex GOFAST, floating point emulation library
5287 @findex gofast_maybe_init_libfuncs
5288 @defmac US_SOFTWARE_GOFAST
5289 Define this macro if your system C library uses the US Software GOFAST
5290 library to provide floating point emulation.
5292 In addition to defining this macro, your architecture must set
5293 @code{TARGET_INIT_LIBFUNCS} to @code{gofast_maybe_init_libfuncs}, or
5294 else call that function from its version of that hook. It is defined
5295 in @file{config/gofast.h}, which must be included by your
5296 architecture's @file{@var{cpu}.c} file. See @file{sparc/sparc.c} for
5299 If this macro is defined, the
5300 @code{TARGET_FLOAT_LIB_COMPARE_RETURNS_BOOL} target hook must return
5301 false for @code{SFmode} and @code{DFmode} comparisons.
5304 @cindex @code{EDOM}, implicit usage
5307 The value of @code{EDOM} on the target machine, as a C integer constant
5308 expression. If you don't define this macro, GCC does not attempt to
5309 deposit the value of @code{EDOM} into @code{errno} directly. Look in
5310 @file{/usr/include/errno.h} to find the value of @code{EDOM} on your
5313 If you do not define @code{TARGET_EDOM}, then compiled code reports
5314 domain errors by calling the library function and letting it report the
5315 error. If mathematical functions on your system use @code{matherr} when
5316 there is an error, then you should leave @code{TARGET_EDOM} undefined so
5317 that @code{matherr} is used normally.
5320 @cindex @code{errno}, implicit usage
5321 @defmac GEN_ERRNO_RTX
5322 Define this macro as a C expression to create an rtl expression that
5323 refers to the global ``variable'' @code{errno}. (On certain systems,
5324 @code{errno} may not actually be a variable.) If you don't define this
5325 macro, a reasonable default is used.
5328 @cindex C99 math functions, implicit usage
5329 @defmac TARGET_C99_FUNCTIONS
5330 When this macro is nonzero, GCC will implicitly optimize @code{sin} calls into
5331 @code{sinf} and similarly for other functions defined by C99 standard. The
5332 default is zero because a number of existing systems lack support for these
5333 functions in their runtime so this macro needs to be redefined to one on
5334 systems that do support the C99 runtime.
5337 @cindex sincos math function, implicit usage
5338 @defmac TARGET_HAS_SINCOS
5339 When this macro is nonzero, GCC will implicitly optimize calls to @code{sin}
5340 and @code{cos} with the same argument to a call to @code{sincos}. The
5341 default is zero. The target has to provide the following functions:
5343 void sincos(double x, double *sin, double *cos);
5344 void sincosf(float x, float *sin, float *cos);
5345 void sincosl(long double x, long double *sin, long double *cos);
5349 @defmac NEXT_OBJC_RUNTIME
5350 Define this macro to generate code for Objective-C message sending using
5351 the calling convention of the NeXT system. This calling convention
5352 involves passing the object, the selector and the method arguments all
5353 at once to the method-lookup library function.
5355 The default calling convention passes just the object and the selector
5356 to the lookup function, which returns a pointer to the method.
5359 @node Addressing Modes
5360 @section Addressing Modes
5361 @cindex addressing modes
5363 @c prevent bad page break with this line
5364 This is about addressing modes.
5366 @defmac HAVE_PRE_INCREMENT
5367 @defmacx HAVE_PRE_DECREMENT
5368 @defmacx HAVE_POST_INCREMENT
5369 @defmacx HAVE_POST_DECREMENT
5370 A C expression that is nonzero if the machine supports pre-increment,
5371 pre-decrement, post-increment, or post-decrement addressing respectively.
5374 @defmac HAVE_PRE_MODIFY_DISP
5375 @defmacx HAVE_POST_MODIFY_DISP
5376 A C expression that is nonzero if the machine supports pre- or
5377 post-address side-effect generation involving constants other than
5378 the size of the memory operand.
5381 @defmac HAVE_PRE_MODIFY_REG
5382 @defmacx HAVE_POST_MODIFY_REG
5383 A C expression that is nonzero if the machine supports pre- or
5384 post-address side-effect generation involving a register displacement.
5387 @defmac CONSTANT_ADDRESS_P (@var{x})
5388 A C expression that is 1 if the RTX @var{x} is a constant which
5389 is a valid address. On most machines the default definition of
5390 @code{(CONSTANT_P (@var{x}) && GET_CODE (@var{x}) != CONST_DOUBLE)}
5391 is acceptable, but a few machines are more restrictive as to which
5392 constant addresses are supported.
5395 @defmac CONSTANT_P (@var{x})
5396 @code{CONSTANT_P}, which is defined by target-independent code,
5397 accepts integer-values expressions whose values are not explicitly
5398 known, such as @code{symbol_ref}, @code{label_ref}, and @code{high}
5399 expressions and @code{const} arithmetic expressions, in addition to
5400 @code{const_int} and @code{const_double} expressions.
5403 @defmac MAX_REGS_PER_ADDRESS
5404 A number, the maximum number of registers that can appear in a valid
5405 memory address. Note that it is up to you to specify a value equal to
5406 the maximum number that @code{TARGET_LEGITIMATE_ADDRESS_P} would ever
5410 @deftypefn {Target Hook} bool TARGET_LEGITIMATE_ADDRESS_P (enum machine_mode @var{mode}, rtx @var{x}, bool @var{strict})
5411 A function that returns whether @var{x} (an RTX) is a legitimate memory
5412 address on the target machine for a memory operand of mode @var{mode}.
5414 Legitimate addresses are defined in two variants: a strict variant and a
5415 non-strict one. The @var{strict} parameter chooses which variant is
5416 desired by the caller.
5418 The strict variant is used in the reload pass. It must be defined so
5419 that any pseudo-register that has not been allocated a hard register is
5420 considered a memory reference. This is because in contexts where some
5421 kind of register is required, a pseudo-register with no hard register
5422 must be rejected. For non-hard registers, the strict variant should look
5423 up the @code{reg_renumber} array; it should then proceed using the hard
5424 register number in the array, or treat the pseudo as a memory reference
5425 if the array holds @code{-1}.
5427 The non-strict variant is used in other passes. It must be defined to
5428 accept all pseudo-registers in every context where some kind of
5429 register is required.
5431 Normally, constant addresses which are the sum of a @code{symbol_ref}
5432 and an integer are stored inside a @code{const} RTX to mark them as
5433 constant. Therefore, there is no need to recognize such sums
5434 specifically as legitimate addresses. Normally you would simply
5435 recognize any @code{const} as legitimate.
5437 Usually @code{PRINT_OPERAND_ADDRESS} is not prepared to handle constant
5438 sums that are not marked with @code{const}. It assumes that a naked
5439 @code{plus} indicates indexing. If so, then you @emph{must} reject such
5440 naked constant sums as illegitimate addresses, so that none of them will
5441 be given to @code{PRINT_OPERAND_ADDRESS}.
5443 @cindex @code{TARGET_ENCODE_SECTION_INFO} and address validation
5444 On some machines, whether a symbolic address is legitimate depends on
5445 the section that the address refers to. On these machines, define the
5446 target hook @code{TARGET_ENCODE_SECTION_INFO} to store the information
5447 into the @code{symbol_ref}, and then check for it here. When you see a
5448 @code{const}, you will have to look inside it to find the
5449 @code{symbol_ref} in order to determine the section. @xref{Assembler
5452 @cindex @code{GO_IF_LEGITIMATE_ADDRESS}
5453 Some ports are still using a deprecated legacy substitute for
5454 this hook, the @code{GO_IF_LEGITIMATE_ADDRESS} macro. This macro
5458 #define GO_IF_LEGITIMATE_ADDRESS (@var{mode}, @var{x}, @var{label})
5462 and should @code{goto @var{label}} if the address @var{x} is a valid
5463 address on the target machine for a memory operand of mode @var{mode}.
5464 Whether the strict or non-strict variants are desired is defined by
5465 the @code{REG_OK_STRICT} macro introduced earlier in this section.
5466 Using the hook is usually simpler because it limits the number of
5467 files that are recompiled when changes are made.
5470 @defmac TARGET_MEM_CONSTRAINT
5471 A single character to be used instead of the default @code{'m'}
5472 character for general memory addresses. This defines the constraint
5473 letter which matches the memory addresses accepted by
5474 @code{TARGET_LEGITIMATE_ADDRESS_P}. Define this macro if you want to
5475 support new address formats in your back end without changing the
5476 semantics of the @code{'m'} constraint. This is necessary in order to
5477 preserve functionality of inline assembly constructs using the
5478 @code{'m'} constraint.
5481 @defmac FIND_BASE_TERM (@var{x})
5482 A C expression to determine the base term of address @var{x},
5483 or to provide a simplified version of @var{x} from which @file{alias.c}
5484 can easily find the base term. This macro is used in only two places:
5485 @code{find_base_value} and @code{find_base_term} in @file{alias.c}.
5487 It is always safe for this macro to not be defined. It exists so
5488 that alias analysis can understand machine-dependent addresses.
5490 The typical use of this macro is to handle addresses containing
5491 a label_ref or symbol_ref within an UNSPEC@.
5494 @deftypefn {Target Hook} rtx TARGET_LEGITIMIZE_ADDRESS (rtx @var{x}, rtx @var{oldx}, enum machine_mode @var{mode})
5495 This hook is given an invalid memory address @var{x} for an
5496 operand of mode @var{mode} and should try to return a valid memory
5499 @findex break_out_memory_refs
5500 @var{x} will always be the result of a call to @code{break_out_memory_refs},
5501 and @var{oldx} will be the operand that was given to that function to produce
5504 The code of the hook should not alter the substructure of
5505 @var{x}. If it transforms @var{x} into a more legitimate form, it
5506 should return the new @var{x}.
5508 It is not necessary for this hook to come up with a legitimate address.
5509 The compiler has standard ways of doing so in all cases. In fact, it
5510 is safe to omit this hook or make it return @var{x} if it cannot find
5511 a valid way to legitimize the address. But often a machine-dependent
5512 strategy can generate better code.
5515 @defmac LEGITIMIZE_RELOAD_ADDRESS (@var{x}, @var{mode}, @var{opnum}, @var{type}, @var{ind_levels}, @var{win})
5516 A C compound statement that attempts to replace @var{x}, which is an address
5517 that needs reloading, with a valid memory address for an operand of mode
5518 @var{mode}. @var{win} will be a C statement label elsewhere in the code.
5519 It is not necessary to define this macro, but it might be useful for
5520 performance reasons.
5522 For example, on the i386, it is sometimes possible to use a single
5523 reload register instead of two by reloading a sum of two pseudo
5524 registers into a register. On the other hand, for number of RISC
5525 processors offsets are limited so that often an intermediate address
5526 needs to be generated in order to address a stack slot. By defining
5527 @code{LEGITIMIZE_RELOAD_ADDRESS} appropriately, the intermediate addresses
5528 generated for adjacent some stack slots can be made identical, and thus
5531 @emph{Note}: This macro should be used with caution. It is necessary
5532 to know something of how reload works in order to effectively use this,
5533 and it is quite easy to produce macros that build in too much knowledge
5534 of reload internals.
5536 @emph{Note}: This macro must be able to reload an address created by a
5537 previous invocation of this macro. If it fails to handle such addresses
5538 then the compiler may generate incorrect code or abort.
5541 The macro definition should use @code{push_reload} to indicate parts that
5542 need reloading; @var{opnum}, @var{type} and @var{ind_levels} are usually
5543 suitable to be passed unaltered to @code{push_reload}.
5545 The code generated by this macro must not alter the substructure of
5546 @var{x}. If it transforms @var{x} into a more legitimate form, it
5547 should assign @var{x} (which will always be a C variable) a new value.
5548 This also applies to parts that you change indirectly by calling
5551 @findex strict_memory_address_p
5552 The macro definition may use @code{strict_memory_address_p} to test if
5553 the address has become legitimate.
5556 If you want to change only a part of @var{x}, one standard way of doing
5557 this is to use @code{copy_rtx}. Note, however, that it unshares only a
5558 single level of rtl. Thus, if the part to be changed is not at the
5559 top level, you'll need to replace first the top level.
5560 It is not necessary for this macro to come up with a legitimate
5561 address; but often a machine-dependent strategy can generate better code.
5564 @deftypefn {Target Hook} bool TARGET_MODE_DEPENDENT_ADDRESS_P (const_rtx @var{addr})
5565 This hook returns @code{true} if memory address @var{addr} can have
5566 different meanings depending on the machine mode of the memory
5567 reference it is used for or if the address is valid for some modes
5570 Autoincrement and autodecrement addresses typically have mode-dependent
5571 effects because the amount of the increment or decrement is the size
5572 of the operand being addressed. Some machines have other mode-dependent
5573 addresses. Many RISC machines have no mode-dependent addresses.
5575 You may assume that @var{addr} is a valid address for the machine.
5577 The default version of this hook returns @code{false}.
5580 @defmac GO_IF_MODE_DEPENDENT_ADDRESS (@var{addr}, @var{label})
5581 A C statement or compound statement with a conditional @code{goto
5582 @var{label};} executed if memory address @var{x} (an RTX) can have
5583 different meanings depending on the machine mode of the memory
5584 reference it is used for or if the address is valid for some modes
5587 Autoincrement and autodecrement addresses typically have mode-dependent
5588 effects because the amount of the increment or decrement is the size
5589 of the operand being addressed. Some machines have other mode-dependent
5590 addresses. Many RISC machines have no mode-dependent addresses.
5592 You may assume that @var{addr} is a valid address for the machine.
5594 These are obsolete macros, replaced by the
5595 @code{TARGET_MODE_DEPENDENT_ADDRESS_P} target hook.
5598 @defmac LEGITIMATE_CONSTANT_P (@var{x})
5599 A C expression that is nonzero if @var{x} is a legitimate constant for
5600 an immediate operand on the target machine. You can assume that
5601 @var{x} satisfies @code{CONSTANT_P}, so you need not check this. In fact,
5602 @samp{1} is a suitable definition for this macro on machines where
5603 anything @code{CONSTANT_P} is valid.
5606 @deftypefn {Target Hook} rtx TARGET_DELEGITIMIZE_ADDRESS (rtx @var{x})
5607 This hook is used to undo the possibly obfuscating effects of the
5608 @code{LEGITIMIZE_ADDRESS} and @code{LEGITIMIZE_RELOAD_ADDRESS} target
5609 macros. Some backend implementations of these macros wrap symbol
5610 references inside an @code{UNSPEC} rtx to represent PIC or similar
5611 addressing modes. This target hook allows GCC's optimizers to understand
5612 the semantics of these opaque @code{UNSPEC}s by converting them back
5613 into their original form.
5616 @deftypefn {Target Hook} bool TARGET_CANNOT_FORCE_CONST_MEM (rtx @var{x})
5617 This hook should return true if @var{x} is of a form that cannot (or
5618 should not) be spilled to the constant pool. The default version of
5619 this hook returns false.
5621 The primary reason to define this hook is to prevent reload from
5622 deciding that a non-legitimate constant would be better reloaded
5623 from the constant pool instead of spilling and reloading a register
5624 holding the constant. This restriction is often true of addresses
5625 of TLS symbols for various targets.
5628 @deftypefn {Target Hook} bool TARGET_USE_BLOCKS_FOR_CONSTANT_P (enum machine_mode @var{mode}, const_rtx @var{x})
5629 This hook should return true if pool entries for constant @var{x} can
5630 be placed in an @code{object_block} structure. @var{mode} is the mode
5633 The default version returns false for all constants.
5636 @deftypefn {Target Hook} tree TARGET_BUILTIN_RECIPROCAL (unsigned @var{fn}, bool @var{md_fn}, bool @var{sqrt})
5637 This hook should return the DECL of a function that implements reciprocal of
5638 the builtin function with builtin function code @var{fn}, or
5639 @code{NULL_TREE} if such a function is not available. @var{md_fn} is true
5640 when @var{fn} is a code of a machine-dependent builtin function. When
5641 @var{sqrt} is true, additional optimizations that apply only to the reciprocal
5642 of a square root function are performed, and only reciprocals of @code{sqrt}
5646 @deftypefn {Target Hook} tree TARGET_VECTORIZE_BUILTIN_MASK_FOR_LOAD (void)
5647 This hook should return the DECL of a function @var{f} that given an
5648 address @var{addr} as an argument returns a mask @var{m} that can be
5649 used to extract from two vectors the relevant data that resides in
5650 @var{addr} in case @var{addr} is not properly aligned.
5652 The autovectorizer, when vectorizing a load operation from an address
5653 @var{addr} that may be unaligned, will generate two vector loads from
5654 the two aligned addresses around @var{addr}. It then generates a
5655 @code{REALIGN_LOAD} operation to extract the relevant data from the
5656 two loaded vectors. The first two arguments to @code{REALIGN_LOAD},
5657 @var{v1} and @var{v2}, are the two vectors, each of size @var{VS}, and
5658 the third argument, @var{OFF}, defines how the data will be extracted
5659 from these two vectors: if @var{OFF} is 0, then the returned vector is
5660 @var{v2}; otherwise, the returned vector is composed from the last
5661 @var{VS}-@var{OFF} elements of @var{v1} concatenated to the first
5662 @var{OFF} elements of @var{v2}.
5664 If this hook is defined, the autovectorizer will generate a call
5665 to @var{f} (using the DECL tree that this hook returns) and will
5666 use the return value of @var{f} as the argument @var{OFF} to
5667 @code{REALIGN_LOAD}. Therefore, the mask @var{m} returned by @var{f}
5668 should comply with the semantics expected by @code{REALIGN_LOAD}
5670 If this hook is not defined, then @var{addr} will be used as
5671 the argument @var{OFF} to @code{REALIGN_LOAD}, in which case the low
5672 log2(@var{VS}) @minus{} 1 bits of @var{addr} will be considered.
5675 @deftypefn {Target Hook} tree TARGET_VECTORIZE_BUILTIN_MUL_WIDEN_EVEN (tree @var{x})
5676 This hook should return the DECL of a function @var{f} that implements
5677 widening multiplication of the even elements of two input vectors of type @var{x}.
5679 If this hook is defined, the autovectorizer will use it along with the
5680 @code{TARGET_VECTORIZE_BUILTIN_MUL_WIDEN_ODD} target hook when vectorizing
5681 widening multiplication in cases that the order of the results does not have to be
5682 preserved (e.g.@: used only by a reduction computation). Otherwise, the
5683 @code{widen_mult_hi/lo} idioms will be used.
5686 @deftypefn {Target Hook} tree TARGET_VECTORIZE_BUILTIN_MUL_WIDEN_ODD (tree @var{x})
5687 This hook should return the DECL of a function @var{f} that implements
5688 widening multiplication of the odd elements of two input vectors of type @var{x}.
5690 If this hook is defined, the autovectorizer will use it along with the
5691 @code{TARGET_VECTORIZE_BUILTIN_MUL_WIDEN_EVEN} target hook when vectorizing
5692 widening multiplication in cases that the order of the results does not have to be
5693 preserved (e.g.@: used only by a reduction computation). Otherwise, the
5694 @code{widen_mult_hi/lo} idioms will be used.
5697 @deftypefn {Target Hook} int TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST (enum vect_cost_for_stmt @var{type_of_cost}, tree @var{vectype}, int @var{misalign})
5698 Returns cost of different scalar or vector statements for vectorization cost model.
5699 For vector memory operations the cost may depend on type (@var{vectype}) and
5700 misalignment value (@var{misalign}).
5703 @deftypefn {Target Hook} bool TARGET_VECTORIZE_VECTOR_ALIGNMENT_REACHABLE (const_tree @var{type}, bool @var{is_packed})
5704 Return true if vector alignment is reachable (by peeling N iterations) for the given type.
5707 @deftypefn {Target Hook} tree TARGET_VECTORIZE_BUILTIN_VEC_PERM (tree @var{type}, tree *@var{mask_element_type})
5708 Target builtin that implements vector permute.
5711 @deftypefn {Target Hook} bool TARGET_VECTORIZE_BUILTIN_VEC_PERM_OK (tree @var{vec_type}, tree @var{mask})
5712 Return true if a vector created for @code{builtin_vec_perm} is valid.
5715 @deftypefn {Target Hook} tree TARGET_VECTORIZE_BUILTIN_CONVERSION (unsigned @var{code}, tree @var{dest_type}, tree @var{src_type})
5716 This hook should return the DECL of a function that implements conversion of the
5717 input vector of type @var{src_type} to type @var{dest_type}.
5718 The value of @var{code} is one of the enumerators in @code{enum tree_code} and
5719 specifies how the conversion is to be applied
5720 (truncation, rounding, etc.).
5722 If this hook is defined, the autovectorizer will use the
5723 @code{TARGET_VECTORIZE_BUILTIN_CONVERSION} target hook when vectorizing
5724 conversion. Otherwise, it will return @code{NULL_TREE}.
5727 @deftypefn {Target Hook} tree TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION (tree @var{fndecl}, tree @var{vec_type_out}, tree @var{vec_type_in})
5728 This hook should return the decl of a function that implements the
5729 vectorized variant of the builtin function with builtin function code
5730 @var{code} or @code{NULL_TREE} if such a function is not available.
5731 The value of @var{fndecl} is the builtin function declaration. The
5732 return type of the vectorized function shall be of vector type
5733 @var{vec_type_out} and the argument types should be @var{vec_type_in}.
5736 @deftypefn {Target Hook} bool TARGET_VECTORIZE_SUPPORT_VECTOR_MISALIGNMENT (enum machine_mode @var{mode}, const_tree @var{type}, int @var{misalignment}, bool @var{is_packed})
5737 This hook should return true if the target supports misaligned vector
5738 store/load of a specific factor denoted in the @var{misalignment}
5739 parameter. The vector store/load should be of machine mode @var{mode} and
5740 the elements in the vectors should be of type @var{type}. @var{is_packed}
5741 parameter is true if the memory access is defined in a packed struct.
5744 @deftypefn {Target Hook} {unsigned int} TARGET_VECTORIZE_UNITS_PER_SIMD_WORD (enum machine_mode @var{mode})
5745 This hook should return th number of units in the vectors that the
5746 vectorizer can produce for scalar mode @var{mode}. The default is
5747 equal to @code{UNITS_PER_WORD}, because the vectorizer can do some
5748 transformations even in absence of specialized @acronym{SIMD} hardware.
5751 @node Anchored Addresses
5752 @section Anchored Addresses
5753 @cindex anchored addresses
5754 @cindex @option{-fsection-anchors}
5756 GCC usually addresses every static object as a separate entity.
5757 For example, if we have:
5761 int foo (void) @{ return a + b + c; @}
5764 the code for @code{foo} will usually calculate three separate symbolic
5765 addresses: those of @code{a}, @code{b} and @code{c}. On some targets,
5766 it would be better to calculate just one symbolic address and access
5767 the three variables relative to it. The equivalent pseudocode would
5773 register int *xr = &x;
5774 return xr[&a - &x] + xr[&b - &x] + xr[&c - &x];
5778 (which isn't valid C). We refer to shared addresses like @code{x} as
5779 ``section anchors''. Their use is controlled by @option{-fsection-anchors}.
5781 The hooks below describe the target properties that GCC needs to know
5782 in order to make effective use of section anchors. It won't use
5783 section anchors at all unless either @code{TARGET_MIN_ANCHOR_OFFSET}
5784 or @code{TARGET_MAX_ANCHOR_OFFSET} is set to a nonzero value.
5786 @deftypevr {Target Hook} HOST_WIDE_INT TARGET_MIN_ANCHOR_OFFSET
5787 The minimum offset that should be applied to a section anchor.
5788 On most targets, it should be the smallest offset that can be
5789 applied to a base register while still giving a legitimate address
5790 for every mode. The default value is 0.
5793 @deftypevr {Target Hook} HOST_WIDE_INT TARGET_MAX_ANCHOR_OFFSET
5794 Like @code{TARGET_MIN_ANCHOR_OFFSET}, but the maximum (inclusive)
5795 offset that should be applied to section anchors. The default
5799 @deftypefn {Target Hook} void TARGET_ASM_OUTPUT_ANCHOR (rtx @var{x})
5800 Write the assembly code to define section anchor @var{x}, which is a
5801 @code{SYMBOL_REF} for which @samp{SYMBOL_REF_ANCHOR_P (@var{x})} is true.
5802 The hook is called with the assembly output position set to the beginning
5803 of @code{SYMBOL_REF_BLOCK (@var{x})}.
5805 If @code{ASM_OUTPUT_DEF} is available, the hook's default definition uses
5806 it to define the symbol as @samp{. + SYMBOL_REF_BLOCK_OFFSET (@var{x})}.
5807 If @code{ASM_OUTPUT_DEF} is not available, the hook's default definition
5808 is @code{NULL}, which disables the use of section anchors altogether.
5811 @deftypefn {Target Hook} bool TARGET_USE_ANCHORS_FOR_SYMBOL_P (const_rtx @var{x})
5812 Return true if GCC should attempt to use anchors to access @code{SYMBOL_REF}
5813 @var{x}. You can assume @samp{SYMBOL_REF_HAS_BLOCK_INFO_P (@var{x})} and
5814 @samp{!SYMBOL_REF_ANCHOR_P (@var{x})}.
5816 The default version is correct for most targets, but you might need to
5817 intercept this hook to handle things like target-specific attributes
5818 or target-specific sections.
5821 @node Condition Code
5822 @section Condition Code Status
5823 @cindex condition code status
5825 The macros in this section can be split in two families, according to the
5826 two ways of representing condition codes in GCC.
5828 The first representation is the so called @code{(cc0)} representation
5829 (@pxref{Jump Patterns}), where all instructions can have an implicit
5830 clobber of the condition codes. The second is the condition code
5831 register representation, which provides better schedulability for
5832 architectures that do have a condition code register, but on which
5833 most instructions do not affect it. The latter category includes
5836 The implicit clobbering poses a strong restriction on the placement of
5837 the definition and use of the condition code, which need to be in adjacent
5838 insns for machines using @code{(cc0)}. This can prevent important
5839 optimizations on some machines. For example, on the IBM RS/6000, there
5840 is a delay for taken branches unless the condition code register is set
5841 three instructions earlier than the conditional branch. The instruction
5842 scheduler cannot perform this optimization if it is not permitted to
5843 separate the definition and use of the condition code register.
5845 For this reason, it is possible and suggested to use a register to
5846 represent the condition code for new ports. If there is a specific
5847 condition code register in the machine, use a hard register. If the
5848 condition code or comparison result can be placed in any general register,
5849 or if there are multiple condition registers, use a pseudo register.
5850 Registers used to store the condition code value will usually have a mode
5851 that is in class @code{MODE_CC}.
5853 Alternatively, you can use @code{BImode} if the comparison operator is
5854 specified already in the compare instruction. In this case, you are not
5855 interested in most macros in this section.
5858 * CC0 Condition Codes:: Old style representation of condition codes.
5859 * MODE_CC Condition Codes:: Modern representation of condition codes.
5860 * Cond. Exec. Macros:: Macros to control conditional execution.
5863 @node CC0 Condition Codes
5864 @subsection Representation of condition codes using @code{(cc0)}
5868 The file @file{conditions.h} defines a variable @code{cc_status} to
5869 describe how the condition code was computed (in case the interpretation of
5870 the condition code depends on the instruction that it was set by). This
5871 variable contains the RTL expressions on which the condition code is
5872 currently based, and several standard flags.
5874 Sometimes additional machine-specific flags must be defined in the machine
5875 description header file. It can also add additional machine-specific
5876 information by defining @code{CC_STATUS_MDEP}.
5878 @defmac CC_STATUS_MDEP
5879 C code for a data type which is used for declaring the @code{mdep}
5880 component of @code{cc_status}. It defaults to @code{int}.
5882 This macro is not used on machines that do not use @code{cc0}.
5885 @defmac CC_STATUS_MDEP_INIT
5886 A C expression to initialize the @code{mdep} field to ``empty''.
5887 The default definition does nothing, since most machines don't use
5888 the field anyway. If you want to use the field, you should probably
5889 define this macro to initialize it.
5891 This macro is not used on machines that do not use @code{cc0}.
5894 @defmac NOTICE_UPDATE_CC (@var{exp}, @var{insn})
5895 A C compound statement to set the components of @code{cc_status}
5896 appropriately for an insn @var{insn} whose body is @var{exp}. It is
5897 this macro's responsibility to recognize insns that set the condition
5898 code as a byproduct of other activity as well as those that explicitly
5901 This macro is not used on machines that do not use @code{cc0}.
5903 If there are insns that do not set the condition code but do alter
5904 other machine registers, this macro must check to see whether they
5905 invalidate the expressions that the condition code is recorded as
5906 reflecting. For example, on the 68000, insns that store in address
5907 registers do not set the condition code, which means that usually
5908 @code{NOTICE_UPDATE_CC} can leave @code{cc_status} unaltered for such
5909 insns. But suppose that the previous insn set the condition code
5910 based on location @samp{a4@@(102)} and the current insn stores a new
5911 value in @samp{a4}. Although the condition code is not changed by
5912 this, it will no longer be true that it reflects the contents of
5913 @samp{a4@@(102)}. Therefore, @code{NOTICE_UPDATE_CC} must alter
5914 @code{cc_status} in this case to say that nothing is known about the
5915 condition code value.
5917 The definition of @code{NOTICE_UPDATE_CC} must be prepared to deal
5918 with the results of peephole optimization: insns whose patterns are
5919 @code{parallel} RTXs containing various @code{reg}, @code{mem} or
5920 constants which are just the operands. The RTL structure of these
5921 insns is not sufficient to indicate what the insns actually do. What
5922 @code{NOTICE_UPDATE_CC} should do when it sees one is just to run
5923 @code{CC_STATUS_INIT}.
5925 A possible definition of @code{NOTICE_UPDATE_CC} is to call a function
5926 that looks at an attribute (@pxref{Insn Attributes}) named, for example,
5927 @samp{cc}. This avoids having detailed information about patterns in
5928 two places, the @file{md} file and in @code{NOTICE_UPDATE_CC}.
5931 @node MODE_CC Condition Codes
5932 @subsection Representation of condition codes using registers
5936 @defmac SELECT_CC_MODE (@var{op}, @var{x}, @var{y})
5937 On many machines, the condition code may be produced by other instructions
5938 than compares, for example the branch can use directly the condition
5939 code set by a subtract instruction. However, on some machines
5940 when the condition code is set this way some bits (such as the overflow
5941 bit) are not set in the same way as a test instruction, so that a different
5942 branch instruction must be used for some conditional branches. When
5943 this happens, use the machine mode of the condition code register to
5944 record different formats of the condition code register. Modes can
5945 also be used to record which compare instruction (e.g. a signed or an
5946 unsigned comparison) produced the condition codes.
5948 If other modes than @code{CCmode} are required, add them to
5949 @file{@var{machine}-modes.def} and define @code{SELECT_CC_MODE} to choose
5950 a mode given an operand of a compare. This is needed because the modes
5951 have to be chosen not only during RTL generation but also, for example,
5952 by instruction combination. The result of @code{SELECT_CC_MODE} should
5953 be consistent with the mode used in the patterns; for example to support
5954 the case of the add on the SPARC discussed above, we have the pattern
5958 [(set (reg:CC_NOOV 0)
5960 (plus:SI (match_operand:SI 0 "register_operand" "%r")
5961 (match_operand:SI 1 "arith_operand" "rI"))
5968 together with a @code{SELECT_CC_MODE} that returns @code{CC_NOOVmode}
5969 for comparisons whose argument is a @code{plus}:
5972 #define SELECT_CC_MODE(OP,X,Y) \
5973 (GET_MODE_CLASS (GET_MODE (X)) == MODE_FLOAT \
5974 ? ((OP == EQ || OP == NE) ? CCFPmode : CCFPEmode) \
5975 : ((GET_CODE (X) == PLUS || GET_CODE (X) == MINUS \
5976 || GET_CODE (X) == NEG) \
5977 ? CC_NOOVmode : CCmode))
5980 Another reason to use modes is to retain information on which operands
5981 were used by the comparison; see @code{REVERSIBLE_CC_MODE} later in
5984 You should define this macro if and only if you define extra CC modes
5985 in @file{@var{machine}-modes.def}.
5988 @defmac CANONICALIZE_COMPARISON (@var{code}, @var{op0}, @var{op1})
5989 On some machines not all possible comparisons are defined, but you can
5990 convert an invalid comparison into a valid one. For example, the Alpha
5991 does not have a @code{GT} comparison, but you can use an @code{LT}
5992 comparison instead and swap the order of the operands.
5994 On such machines, define this macro to be a C statement to do any
5995 required conversions. @var{code} is the initial comparison code
5996 and @var{op0} and @var{op1} are the left and right operands of the
5997 comparison, respectively. You should modify @var{code}, @var{op0}, and
5998 @var{op1} as required.
6000 GCC will not assume that the comparison resulting from this macro is
6001 valid but will see if the resulting insn matches a pattern in the
6004 You need not define this macro if it would never change the comparison
6008 @defmac REVERSIBLE_CC_MODE (@var{mode})
6009 A C expression whose value is one if it is always safe to reverse a
6010 comparison whose mode is @var{mode}. If @code{SELECT_CC_MODE}
6011 can ever return @var{mode} for a floating-point inequality comparison,
6012 then @code{REVERSIBLE_CC_MODE (@var{mode})} must be zero.
6014 You need not define this macro if it would always returns zero or if the
6015 floating-point format is anything other than @code{IEEE_FLOAT_FORMAT}.
6016 For example, here is the definition used on the SPARC, where floating-point
6017 inequality comparisons are always given @code{CCFPEmode}:
6020 #define REVERSIBLE_CC_MODE(MODE) ((MODE) != CCFPEmode)
6024 @defmac REVERSE_CONDITION (@var{code}, @var{mode})
6025 A C expression whose value is reversed condition code of the @var{code} for
6026 comparison done in CC_MODE @var{mode}. The macro is used only in case
6027 @code{REVERSIBLE_CC_MODE (@var{mode})} is nonzero. Define this macro in case
6028 machine has some non-standard way how to reverse certain conditionals. For
6029 instance in case all floating point conditions are non-trapping, compiler may
6030 freely convert unordered compares to ordered one. Then definition may look
6034 #define REVERSE_CONDITION(CODE, MODE) \
6035 ((MODE) != CCFPmode ? reverse_condition (CODE) \
6036 : reverse_condition_maybe_unordered (CODE))
6040 @deftypefn {Target Hook} bool TARGET_FIXED_CONDITION_CODE_REGS (unsigned int *@var{p1}, unsigned int *@var{p2})
6041 On targets which do not use @code{(cc0)}, and which use a hard
6042 register rather than a pseudo-register to hold condition codes, the
6043 regular CSE passes are often not able to identify cases in which the
6044 hard register is set to a common value. Use this hook to enable a
6045 small pass which optimizes such cases. This hook should return true
6046 to enable this pass, and it should set the integers to which its
6047 arguments point to the hard register numbers used for condition codes.
6048 When there is only one such register, as is true on most systems, the
6049 integer pointed to by @var{p2} should be set to
6050 @code{INVALID_REGNUM}.
6052 The default version of this hook returns false.
6055 @deftypefn {Target Hook} {enum machine_mode} TARGET_CC_MODES_COMPATIBLE (enum machine_mode @var{m1}, enum machine_mode @var{m2})
6056 On targets which use multiple condition code modes in class
6057 @code{MODE_CC}, it is sometimes the case that a comparison can be
6058 validly done in more than one mode. On such a system, define this
6059 target hook to take two mode arguments and to return a mode in which
6060 both comparisons may be validly done. If there is no such mode,
6061 return @code{VOIDmode}.
6063 The default version of this hook checks whether the modes are the
6064 same. If they are, it returns that mode. If they are different, it
6065 returns @code{VOIDmode}.
6068 @node Cond. Exec. Macros
6069 @subsection Macros to control conditional execution
6070 @findex conditional execution
6073 There is one macro that may need to be defined for targets
6074 supporting conditional execution, independent of how they
6075 represent conditional branches.
6077 @defmac REVERSE_CONDEXEC_PREDICATES_P (@var{op1}, @var{op2})
6078 A C expression that returns true if the conditional execution predicate
6079 @var{op1}, a comparison operation, is the inverse of @var{op2} and vice
6080 versa. Define this to return 0 if the target has conditional execution
6081 predicates that cannot be reversed safely. There is no need to validate
6082 that the arguments of op1 and op2 are the same, this is done separately.
6083 If no expansion is specified, this macro is defined as follows:
6086 #define REVERSE_CONDEXEC_PREDICATES_P (x, y) \
6087 (GET_CODE ((x)) == reversed_comparison_code ((y), NULL))
6092 @section Describing Relative Costs of Operations
6093 @cindex costs of instructions
6094 @cindex relative costs
6095 @cindex speed of instructions
6097 These macros let you describe the relative speed of various operations
6098 on the target machine.
6100 @defmac REGISTER_MOVE_COST (@var{mode}, @var{from}, @var{to})
6101 A C expression for the cost of moving data of mode @var{mode} from a
6102 register in class @var{from} to one in class @var{to}. The classes are
6103 expressed using the enumeration values such as @code{GENERAL_REGS}. A
6104 value of 2 is the default; other values are interpreted relative to
6107 It is not required that the cost always equal 2 when @var{from} is the
6108 same as @var{to}; on some machines it is expensive to move between
6109 registers if they are not general registers.
6111 If reload sees an insn consisting of a single @code{set} between two
6112 hard registers, and if @code{REGISTER_MOVE_COST} applied to their
6113 classes returns a value of 2, reload does not check to ensure that the
6114 constraints of the insn are met. Setting a cost of other than 2 will
6115 allow reload to verify that the constraints are met. You should do this
6116 if the @samp{mov@var{m}} pattern's constraints do not allow such copying.
6118 These macros are obsolete, new ports should use the target hook
6119 @code{TARGET_REGISTER_MOVE_COST} instead.
6122 @deftypefn {Target Hook} int TARGET_REGISTER_MOVE_COST (enum machine_mode @var{mode}, reg_class_t @var{from}, reg_class_t @var{to})
6123 This target hook should return the cost of moving data of mode @var{mode}
6124 from a register in class @var{from} to one in class @var{to}. The classes
6125 are expressed using the enumeration values such as @code{GENERAL_REGS}.
6126 A value of 2 is the default; other values are interpreted relative to
6129 It is not required that the cost always equal 2 when @var{from} is the
6130 same as @var{to}; on some machines it is expensive to move between
6131 registers if they are not general registers.
6133 If reload sees an insn consisting of a single @code{set} between two
6134 hard registers, and if @code{TARGET_REGISTER_MOVE_COST} applied to their
6135 classes returns a value of 2, reload does not check to ensure that the
6136 constraints of the insn are met. Setting a cost of other than 2 will
6137 allow reload to verify that the constraints are met. You should do this
6138 if the @samp{mov@var{m}} pattern's constraints do not allow such copying.
6140 The default version of this function returns 2.
6143 @defmac MEMORY_MOVE_COST (@var{mode}, @var{class}, @var{in})
6144 A C expression for the cost of moving data of mode @var{mode} between a
6145 register of class @var{class} and memory; @var{in} is zero if the value
6146 is to be written to memory, nonzero if it is to be read in. This cost
6147 is relative to those in @code{REGISTER_MOVE_COST}. If moving between
6148 registers and memory is more expensive than between two registers, you
6149 should define this macro to express the relative cost.
6151 If you do not define this macro, GCC uses a default cost of 4 plus
6152 the cost of copying via a secondary reload register, if one is
6153 needed. If your machine requires a secondary reload register to copy
6154 between memory and a register of @var{class} but the reload mechanism is
6155 more complex than copying via an intermediate, define this macro to
6156 reflect the actual cost of the move.
6158 GCC defines the function @code{memory_move_secondary_cost} if
6159 secondary reloads are needed. It computes the costs due to copying via
6160 a secondary register. If your machine copies from memory using a
6161 secondary register in the conventional way but the default base value of
6162 4 is not correct for your machine, define this macro to add some other
6163 value to the result of that function. The arguments to that function
6164 are the same as to this macro.
6166 These macros are obsolete, new ports should use the target hook
6167 @code{TARGET_MEMORY_MOVE_COST} instead.
6170 @deftypefn {Target Hook} int TARGET_MEMORY_MOVE_COST (enum machine_mode @var{mode}, reg_class_t @var{rclass}, bool @var{in})
6171 This target hook should return the cost of moving data of mode @var{mode}
6172 between a register of class @var{rclass} and memory; @var{in} is @code{false}
6173 if the value is to be written to memory, @code{true} if it is to be read in.
6174 This cost is relative to those in @code{TARGET_REGISTER_MOVE_COST}.
6175 If moving between registers and memory is more expensive than between two
6176 registers, you should add this target hook to express the relative cost.
6178 If you do not add this target hook, GCC uses a default cost of 4 plus
6179 the cost of copying via a secondary reload register, if one is
6180 needed. If your machine requires a secondary reload register to copy
6181 between memory and a register of @var{rclass} but the reload mechanism is
6182 more complex than copying via an intermediate, use this target hook to
6183 reflect the actual cost of the move.
6185 GCC defines the function @code{memory_move_secondary_cost} if
6186 secondary reloads are needed. It computes the costs due to copying via
6187 a secondary register. If your machine copies from memory using a
6188 secondary register in the conventional way but the default base value of
6189 4 is not correct for your machine, use this target hook to add some other
6190 value to the result of that function. The arguments to that function
6191 are the same as to this target hook.
6194 @defmac BRANCH_COST (@var{speed_p}, @var{predictable_p})
6195 A C expression for the cost of a branch instruction. A value of 1 is the
6196 default; other values are interpreted relative to that. Parameter @var{speed_p}
6197 is true when the branch in question should be optimized for speed. When
6198 it is false, @code{BRANCH_COST} should be returning value optimal for code size
6199 rather then performance considerations. @var{predictable_p} is true for well
6200 predictable branches. On many architectures the @code{BRANCH_COST} can be
6204 Here are additional macros which do not specify precise relative costs,
6205 but only that certain actions are more expensive than GCC would
6208 @defmac SLOW_BYTE_ACCESS
6209 Define this macro as a C expression which is nonzero if accessing less
6210 than a word of memory (i.e.@: a @code{char} or a @code{short}) is no
6211 faster than accessing a word of memory, i.e., if such access
6212 require more than one instruction or if there is no difference in cost
6213 between byte and (aligned) word loads.
6215 When this macro is not defined, the compiler will access a field by
6216 finding the smallest containing object; when it is defined, a fullword
6217 load will be used if alignment permits. Unless bytes accesses are
6218 faster than word accesses, using word accesses is preferable since it
6219 may eliminate subsequent memory access if subsequent accesses occur to
6220 other fields in the same word of the structure, but to different bytes.
6223 @defmac SLOW_UNALIGNED_ACCESS (@var{mode}, @var{alignment})
6224 Define this macro to be the value 1 if memory accesses described by the
6225 @var{mode} and @var{alignment} parameters have a cost many times greater
6226 than aligned accesses, for example if they are emulated in a trap
6229 When this macro is nonzero, the compiler will act as if
6230 @code{STRICT_ALIGNMENT} were nonzero when generating code for block
6231 moves. This can cause significantly more instructions to be produced.
6232 Therefore, do not set this macro nonzero if unaligned accesses only add a
6233 cycle or two to the time for a memory access.
6235 If the value of this macro is always zero, it need not be defined. If
6236 this macro is defined, it should produce a nonzero value when
6237 @code{STRICT_ALIGNMENT} is nonzero.
6240 @defmac MOVE_RATIO (@var{speed})
6241 The threshold of number of scalar memory-to-memory move insns, @emph{below}
6242 which a sequence of insns should be generated instead of a
6243 string move insn or a library call. Increasing the value will always
6244 make code faster, but eventually incurs high cost in increased code size.
6246 Note that on machines where the corresponding move insn is a
6247 @code{define_expand} that emits a sequence of insns, this macro counts
6248 the number of such sequences.
6250 The parameter @var{speed} is true if the code is currently being
6251 optimized for speed rather than size.
6253 If you don't define this, a reasonable default is used.
6256 @defmac MOVE_BY_PIECES_P (@var{size}, @var{alignment})
6257 A C expression used to determine whether @code{move_by_pieces} will be used to
6258 copy a chunk of memory, or whether some other block move mechanism
6259 will be used. Defaults to 1 if @code{move_by_pieces_ninsns} returns less
6260 than @code{MOVE_RATIO}.
6263 @defmac MOVE_MAX_PIECES
6264 A C expression used by @code{move_by_pieces} to determine the largest unit
6265 a load or store used to copy memory is. Defaults to @code{MOVE_MAX}.
6268 @defmac CLEAR_RATIO (@var{speed})
6269 The threshold of number of scalar move insns, @emph{below} which a sequence
6270 of insns should be generated to clear memory instead of a string clear insn
6271 or a library call. Increasing the value will always make code faster, but
6272 eventually incurs high cost in increased code size.
6274 The parameter @var{speed} is true if the code is currently being
6275 optimized for speed rather than size.
6277 If you don't define this, a reasonable default is used.
6280 @defmac CLEAR_BY_PIECES_P (@var{size}, @var{alignment})
6281 A C expression used to determine whether @code{clear_by_pieces} will be used
6282 to clear a chunk of memory, or whether some other block clear mechanism
6283 will be used. Defaults to 1 if @code{move_by_pieces_ninsns} returns less
6284 than @code{CLEAR_RATIO}.
6287 @defmac SET_RATIO (@var{speed})
6288 The threshold of number of scalar move insns, @emph{below} which a sequence
6289 of insns should be generated to set memory to a constant value, instead of
6290 a block set insn or a library call.
6291 Increasing the value will always make code faster, but
6292 eventually incurs high cost in increased code size.
6294 The parameter @var{speed} is true if the code is currently being
6295 optimized for speed rather than size.
6297 If you don't define this, it defaults to the value of @code{MOVE_RATIO}.
6300 @defmac SET_BY_PIECES_P (@var{size}, @var{alignment})
6301 A C expression used to determine whether @code{store_by_pieces} will be
6302 used to set a chunk of memory to a constant value, or whether some
6303 other mechanism will be used. Used by @code{__builtin_memset} when
6304 storing values other than constant zero.
6305 Defaults to 1 if @code{move_by_pieces_ninsns} returns less
6306 than @code{SET_RATIO}.
6309 @defmac STORE_BY_PIECES_P (@var{size}, @var{alignment})
6310 A C expression used to determine whether @code{store_by_pieces} will be
6311 used to set a chunk of memory to a constant string value, or whether some
6312 other mechanism will be used. Used by @code{__builtin_strcpy} when
6313 called with a constant source string.
6314 Defaults to 1 if @code{move_by_pieces_ninsns} returns less
6315 than @code{MOVE_RATIO}.
6318 @defmac USE_LOAD_POST_INCREMENT (@var{mode})
6319 A C expression used to determine whether a load postincrement is a good
6320 thing to use for a given mode. Defaults to the value of
6321 @code{HAVE_POST_INCREMENT}.
6324 @defmac USE_LOAD_POST_DECREMENT (@var{mode})
6325 A C expression used to determine whether a load postdecrement is a good
6326 thing to use for a given mode. Defaults to the value of
6327 @code{HAVE_POST_DECREMENT}.
6330 @defmac USE_LOAD_PRE_INCREMENT (@var{mode})
6331 A C expression used to determine whether a load preincrement is a good
6332 thing to use for a given mode. Defaults to the value of
6333 @code{HAVE_PRE_INCREMENT}.
6336 @defmac USE_LOAD_PRE_DECREMENT (@var{mode})
6337 A C expression used to determine whether a load predecrement is a good
6338 thing to use for a given mode. Defaults to the value of
6339 @code{HAVE_PRE_DECREMENT}.
6342 @defmac USE_STORE_POST_INCREMENT (@var{mode})
6343 A C expression used to determine whether a store postincrement is a good
6344 thing to use for a given mode. Defaults to the value of
6345 @code{HAVE_POST_INCREMENT}.
6348 @defmac USE_STORE_POST_DECREMENT (@var{mode})
6349 A C expression used to determine whether a store postdecrement is a good
6350 thing to use for a given mode. Defaults to the value of
6351 @code{HAVE_POST_DECREMENT}.
6354 @defmac USE_STORE_PRE_INCREMENT (@var{mode})
6355 This macro is used to determine whether a store preincrement is a good
6356 thing to use for a given mode. Defaults to the value of
6357 @code{HAVE_PRE_INCREMENT}.
6360 @defmac USE_STORE_PRE_DECREMENT (@var{mode})
6361 This macro is used to determine whether a store predecrement is a good
6362 thing to use for a given mode. Defaults to the value of
6363 @code{HAVE_PRE_DECREMENT}.
6366 @defmac NO_FUNCTION_CSE
6367 Define this macro if it is as good or better to call a constant
6368 function address than to call an address kept in a register.
6371 @defmac RANGE_TEST_NON_SHORT_CIRCUIT
6372 Define this macro if a non-short-circuit operation produced by
6373 @samp{fold_range_test ()} is optimal. This macro defaults to true if
6374 @code{BRANCH_COST} is greater than or equal to the value 2.
6377 @deftypefn {Target Hook} bool TARGET_RTX_COSTS (rtx @var{x}, int @var{code}, int @var{outer_code}, int *@var{total}, bool @var{speed})
6378 This target hook describes the relative costs of RTL expressions.
6380 The cost may depend on the precise form of the expression, which is
6381 available for examination in @var{x}, and the rtx code of the expression
6382 in which it is contained, found in @var{outer_code}. @var{code} is the
6383 expression code---redundant, since it can be obtained with
6384 @code{GET_CODE (@var{x})}.
6386 In implementing this hook, you can use the construct
6387 @code{COSTS_N_INSNS (@var{n})} to specify a cost equal to @var{n} fast
6390 On entry to the hook, @code{*@var{total}} contains a default estimate
6391 for the cost of the expression. The hook should modify this value as
6392 necessary. Traditionally, the default costs are @code{COSTS_N_INSNS (5)}
6393 for multiplications, @code{COSTS_N_INSNS (7)} for division and modulus
6394 operations, and @code{COSTS_N_INSNS (1)} for all other operations.
6396 When optimizing for code size, i.e.@: when @code{speed} is
6397 false, this target hook should be used to estimate the relative
6398 size cost of an expression, again relative to @code{COSTS_N_INSNS}.
6400 The hook returns true when all subexpressions of @var{x} have been
6401 processed, and false when @code{rtx_cost} should recurse.
6404 @deftypefn {Target Hook} int TARGET_ADDRESS_COST (rtx @var{address}, bool @var{speed})
6405 This hook computes the cost of an addressing mode that contains
6406 @var{address}. If not defined, the cost is computed from
6407 the @var{address} expression and the @code{TARGET_RTX_COST} hook.
6409 For most CISC machines, the default cost is a good approximation of the
6410 true cost of the addressing mode. However, on RISC machines, all
6411 instructions normally have the same length and execution time. Hence
6412 all addresses will have equal costs.
6414 In cases where more than one form of an address is known, the form with
6415 the lowest cost will be used. If multiple forms have the same, lowest,
6416 cost, the one that is the most complex will be used.
6418 For example, suppose an address that is equal to the sum of a register
6419 and a constant is used twice in the same basic block. When this macro
6420 is not defined, the address will be computed in a register and memory
6421 references will be indirect through that register. On machines where
6422 the cost of the addressing mode containing the sum is no higher than
6423 that of a simple indirect reference, this will produce an additional
6424 instruction and possibly require an additional register. Proper
6425 specification of this macro eliminates this overhead for such machines.
6427 This hook is never called with an invalid address.
6429 On machines where an address involving more than one register is as
6430 cheap as an address computation involving only one register, defining
6431 @code{TARGET_ADDRESS_COST} to reflect this can cause two registers to
6432 be live over a region of code where only one would have been if
6433 @code{TARGET_ADDRESS_COST} were not defined in that manner. This effect
6434 should be considered in the definition of this macro. Equivalent costs
6435 should probably only be given to addresses with different numbers of
6436 registers on machines with lots of registers.
6440 @section Adjusting the Instruction Scheduler
6442 The instruction scheduler may need a fair amount of machine-specific
6443 adjustment in order to produce good code. GCC provides several target
6444 hooks for this purpose. It is usually enough to define just a few of
6445 them: try the first ones in this list first.
6447 @deftypefn {Target Hook} int TARGET_SCHED_ISSUE_RATE (void)
6448 This hook returns the maximum number of instructions that can ever
6449 issue at the same time on the target machine. The default is one.
6450 Although the insn scheduler can define itself the possibility of issue
6451 an insn on the same cycle, the value can serve as an additional
6452 constraint to issue insns on the same simulated processor cycle (see
6453 hooks @samp{TARGET_SCHED_REORDER} and @samp{TARGET_SCHED_REORDER2}).
6454 This value must be constant over the entire compilation. If you need
6455 it to vary depending on what the instructions are, you must use
6456 @samp{TARGET_SCHED_VARIABLE_ISSUE}.
6459 @deftypefn {Target Hook} int TARGET_SCHED_VARIABLE_ISSUE (FILE *@var{file}, int @var{verbose}, rtx @var{insn}, int @var{more})
6460 This hook is executed by the scheduler after it has scheduled an insn
6461 from the ready list. It should return the number of insns which can
6462 still be issued in the current cycle. The default is
6463 @samp{@w{@var{more} - 1}} for insns other than @code{CLOBBER} and
6464 @code{USE}, which normally are not counted against the issue rate.
6465 You should define this hook if some insns take more machine resources
6466 than others, so that fewer insns can follow them in the same cycle.
6467 @var{file} is either a null pointer, or a stdio stream to write any
6468 debug output to. @var{verbose} is the verbose level provided by
6469 @option{-fsched-verbose-@var{n}}. @var{insn} is the instruction that
6473 @deftypefn {Target Hook} int TARGET_SCHED_ADJUST_COST (rtx @var{insn}, rtx @var{link}, rtx @var{dep_insn}, int @var{cost})
6474 This function corrects the value of @var{cost} based on the
6475 relationship between @var{insn} and @var{dep_insn} through the
6476 dependence @var{link}. It should return the new value. The default
6477 is to make no adjustment to @var{cost}. This can be used for example
6478 to specify to the scheduler using the traditional pipeline description
6479 that an output- or anti-dependence does not incur the same cost as a
6480 data-dependence. If the scheduler using the automaton based pipeline
6481 description, the cost of anti-dependence is zero and the cost of
6482 output-dependence is maximum of one and the difference of latency
6483 times of the first and the second insns. If these values are not
6484 acceptable, you could use the hook to modify them too. See also
6485 @pxref{Processor pipeline description}.
6488 @deftypefn {Target Hook} int TARGET_SCHED_ADJUST_PRIORITY (rtx @var{insn}, int @var{priority})
6489 This hook adjusts the integer scheduling priority @var{priority} of
6490 @var{insn}. It should return the new priority. Increase the priority to
6491 execute @var{insn} earlier, reduce the priority to execute @var{insn}
6492 later. Do not define this hook if you do not need to adjust the
6493 scheduling priorities of insns.
6496 @deftypefn {Target Hook} int TARGET_SCHED_REORDER (FILE *@var{file}, int @var{verbose}, rtx *@var{ready}, int *@var{n_readyp}, int @var{clock})
6497 This hook is executed by the scheduler after it has scheduled the ready
6498 list, to allow the machine description to reorder it (for example to
6499 combine two small instructions together on @samp{VLIW} machines).
6500 @var{file} is either a null pointer, or a stdio stream to write any
6501 debug output to. @var{verbose} is the verbose level provided by
6502 @option{-fsched-verbose-@var{n}}. @var{ready} is a pointer to the ready
6503 list of instructions that are ready to be scheduled. @var{n_readyp} is
6504 a pointer to the number of elements in the ready list. The scheduler
6505 reads the ready list in reverse order, starting with
6506 @var{ready}[@var{*n_readyp} @minus{} 1] and going to @var{ready}[0]. @var{clock}
6507 is the timer tick of the scheduler. You may modify the ready list and
6508 the number of ready insns. The return value is the number of insns that
6509 can issue this cycle; normally this is just @code{issue_rate}. See also
6510 @samp{TARGET_SCHED_REORDER2}.
6513 @deftypefn {Target Hook} int TARGET_SCHED_REORDER2 (FILE *@var{file}, int @var{verbose}, rtx *@var{ready}, int *@var{n_readyp}, int @var{clock})
6514 Like @samp{TARGET_SCHED_REORDER}, but called at a different time. That
6515 function is called whenever the scheduler starts a new cycle. This one
6516 is called once per iteration over a cycle, immediately after
6517 @samp{TARGET_SCHED_VARIABLE_ISSUE}; it can reorder the ready list and
6518 return the number of insns to be scheduled in the same cycle. Defining
6519 this hook can be useful if there are frequent situations where
6520 scheduling one insn causes other insns to become ready in the same
6521 cycle. These other insns can then be taken into account properly.
6524 @deftypefn {Target Hook} void TARGET_SCHED_DEPENDENCIES_EVALUATION_HOOK (rtx @var{head}, rtx @var{tail})
6525 This hook is called after evaluation forward dependencies of insns in
6526 chain given by two parameter values (@var{head} and @var{tail}
6527 correspondingly) but before insns scheduling of the insn chain. For
6528 example, it can be used for better insn classification if it requires
6529 analysis of dependencies. This hook can use backward and forward
6530 dependencies of the insn scheduler because they are already
6534 @deftypefn {Target Hook} void TARGET_SCHED_INIT (FILE *@var{file}, int @var{verbose}, int @var{max_ready})
6535 This hook is executed by the scheduler at the beginning of each block of
6536 instructions that are to be scheduled. @var{file} is either a null
6537 pointer, or a stdio stream to write any debug output to. @var{verbose}
6538 is the verbose level provided by @option{-fsched-verbose-@var{n}}.
6539 @var{max_ready} is the maximum number of insns in the current scheduling
6540 region that can be live at the same time. This can be used to allocate
6541 scratch space if it is needed, e.g.@: by @samp{TARGET_SCHED_REORDER}.
6544 @deftypefn {Target Hook} void TARGET_SCHED_FINISH (FILE *@var{file}, int @var{verbose})
6545 This hook is executed by the scheduler at the end of each block of
6546 instructions that are to be scheduled. It can be used to perform
6547 cleanup of any actions done by the other scheduling hooks. @var{file}
6548 is either a null pointer, or a stdio stream to write any debug output
6549 to. @var{verbose} is the verbose level provided by
6550 @option{-fsched-verbose-@var{n}}.
6553 @deftypefn {Target Hook} void TARGET_SCHED_INIT_GLOBAL (FILE *@var{file}, int @var{verbose}, int @var{old_max_uid})
6554 This hook is executed by the scheduler after function level initializations.
6555 @var{file} is either a null pointer, or a stdio stream to write any debug output to.
6556 @var{verbose} is the verbose level provided by @option{-fsched-verbose-@var{n}}.
6557 @var{old_max_uid} is the maximum insn uid when scheduling begins.
6560 @deftypefn {Target Hook} void TARGET_SCHED_FINISH_GLOBAL (FILE *@var{file}, int @var{verbose})
6561 This is the cleanup hook corresponding to @code{TARGET_SCHED_INIT_GLOBAL}.
6562 @var{file} is either a null pointer, or a stdio stream to write any debug output to.
6563 @var{verbose} is the verbose level provided by @option{-fsched-verbose-@var{n}}.
6566 @deftypefn {Target Hook} rtx TARGET_SCHED_DFA_PRE_CYCLE_INSN (void)
6567 The hook returns an RTL insn. The automaton state used in the
6568 pipeline hazard recognizer is changed as if the insn were scheduled
6569 when the new simulated processor cycle starts. Usage of the hook may
6570 simplify the automaton pipeline description for some @acronym{VLIW}
6571 processors. If the hook is defined, it is used only for the automaton
6572 based pipeline description. The default is not to change the state
6573 when the new simulated processor cycle starts.
6576 @deftypefn {Target Hook} void TARGET_SCHED_INIT_DFA_PRE_CYCLE_INSN (void)
6577 The hook can be used to initialize data used by the previous hook.
6580 @deftypefn {Target Hook} rtx TARGET_SCHED_DFA_POST_CYCLE_INSN (void)
6581 The hook is analogous to @samp{TARGET_SCHED_DFA_PRE_CYCLE_INSN} but used
6582 to changed the state as if the insn were scheduled when the new
6583 simulated processor cycle finishes.
6586 @deftypefn {Target Hook} void TARGET_SCHED_INIT_DFA_POST_CYCLE_INSN (void)
6587 The hook is analogous to @samp{TARGET_SCHED_INIT_DFA_PRE_CYCLE_INSN} but
6588 used to initialize data used by the previous hook.
6591 @deftypefn {Target Hook} void TARGET_SCHED_DFA_PRE_ADVANCE_CYCLE (void)
6592 The hook to notify target that the current simulated cycle is about to finish.
6593 The hook is analogous to @samp{TARGET_SCHED_DFA_PRE_CYCLE_INSN} but used
6594 to change the state in more complicated situations - e.g., when advancing
6595 state on a single insn is not enough.
6598 @deftypefn {Target Hook} void TARGET_SCHED_DFA_POST_ADVANCE_CYCLE (void)
6599 The hook to notify target that new simulated cycle has just started.
6600 The hook is analogous to @samp{TARGET_SCHED_DFA_POST_CYCLE_INSN} but used
6601 to change the state in more complicated situations - e.g., when advancing
6602 state on a single insn is not enough.
6605 @deftypefn {Target Hook} int TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD (void)
6606 This hook controls better choosing an insn from the ready insn queue
6607 for the @acronym{DFA}-based insn scheduler. Usually the scheduler
6608 chooses the first insn from the queue. If the hook returns a positive
6609 value, an additional scheduler code tries all permutations of
6610 @samp{TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD ()}
6611 subsequent ready insns to choose an insn whose issue will result in
6612 maximal number of issued insns on the same cycle. For the
6613 @acronym{VLIW} processor, the code could actually solve the problem of
6614 packing simple insns into the @acronym{VLIW} insn. Of course, if the
6615 rules of @acronym{VLIW} packing are described in the automaton.
6617 This code also could be used for superscalar @acronym{RISC}
6618 processors. Let us consider a superscalar @acronym{RISC} processor
6619 with 3 pipelines. Some insns can be executed in pipelines @var{A} or
6620 @var{B}, some insns can be executed only in pipelines @var{B} or
6621 @var{C}, and one insn can be executed in pipeline @var{B}. The
6622 processor may issue the 1st insn into @var{A} and the 2nd one into
6623 @var{B}. In this case, the 3rd insn will wait for freeing @var{B}
6624 until the next cycle. If the scheduler issues the 3rd insn the first,
6625 the processor could issue all 3 insns per cycle.
6627 Actually this code demonstrates advantages of the automaton based
6628 pipeline hazard recognizer. We try quickly and easy many insn
6629 schedules to choose the best one.
6631 The default is no multipass scheduling.
6634 @deftypefn {Target Hook} int TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD (rtx @var{insn})
6636 This hook controls what insns from the ready insn queue will be
6637 considered for the multipass insn scheduling. If the hook returns
6638 zero for @var{insn}, the insn will be not chosen to
6641 The default is that any ready insns can be chosen to be issued.
6644 @deftypefn {Target Hook} int TARGET_SCHED_DFA_NEW_CYCLE (FILE *@var{dump}, int @var{verbose}, rtx @var{insn}, int @var{last_clock}, int @var{clock}, int *@var{sort_p})
6645 This hook is called by the insn scheduler before issuing @var{insn}
6646 on cycle @var{clock}. If the hook returns nonzero,
6647 @var{insn} is not issued on this processor cycle. Instead,
6648 the processor cycle is advanced. If *@var{sort_p}
6649 is zero, the insn ready queue is not sorted on the new cycle
6650 start as usually. @var{dump} and @var{verbose} specify the file and
6651 verbosity level to use for debugging output.
6652 @var{last_clock} and @var{clock} are, respectively, the
6653 processor cycle on which the previous insn has been issued,
6654 and the current processor cycle.
6657 @deftypefn {Target Hook} bool TARGET_SCHED_IS_COSTLY_DEPENDENCE (struct _dep *@var{_dep}, int @var{cost}, int @var{distance})
6658 This hook is used to define which dependences are considered costly by
6659 the target, so costly that it is not advisable to schedule the insns that
6660 are involved in the dependence too close to one another. The parameters
6661 to this hook are as follows: The first parameter @var{_dep} is the dependence
6662 being evaluated. The second parameter @var{cost} is the cost of the
6663 dependence as estimated by the scheduler, and the third
6664 parameter @var{distance} is the distance in cycles between the two insns.
6665 The hook returns @code{true} if considering the distance between the two
6666 insns the dependence between them is considered costly by the target,
6667 and @code{false} otherwise.
6669 Defining this hook can be useful in multiple-issue out-of-order machines,
6670 where (a) it's practically hopeless to predict the actual data/resource
6671 delays, however: (b) there's a better chance to predict the actual grouping
6672 that will be formed, and (c) correctly emulating the grouping can be very
6673 important. In such targets one may want to allow issuing dependent insns
6674 closer to one another---i.e., closer than the dependence distance; however,
6675 not in cases of ``costly dependences'', which this hooks allows to define.
6678 @deftypefn {Target Hook} void TARGET_SCHED_H_I_D_EXTENDED (void)
6679 This hook is called by the insn scheduler after emitting a new instruction to
6680 the instruction stream. The hook notifies a target backend to extend its
6681 per instruction data structures.
6684 @deftypefn {Target Hook} {void *} TARGET_SCHED_ALLOC_SCHED_CONTEXT (void)
6685 Return a pointer to a store large enough to hold target scheduling context.
6688 @deftypefn {Target Hook} void TARGET_SCHED_INIT_SCHED_CONTEXT (void *@var{tc}, bool @var{clean_p})
6689 Initialize store pointed to by @var{tc} to hold target scheduling context.
6690 It @var{clean_p} is true then initialize @var{tc} as if scheduler is at the
6691 beginning of the block. Otherwise, copy the current context into @var{tc}.
6694 @deftypefn {Target Hook} void TARGET_SCHED_SET_SCHED_CONTEXT (void *@var{tc})
6695 Copy target scheduling context pointed to by @var{tc} to the current context.
6698 @deftypefn {Target Hook} void TARGET_SCHED_CLEAR_SCHED_CONTEXT (void *@var{tc})
6699 Deallocate internal data in target scheduling context pointed to by @var{tc}.
6702 @deftypefn {Target Hook} void TARGET_SCHED_FREE_SCHED_CONTEXT (void *@var{tc})
6703 Deallocate a store for target scheduling context pointed to by @var{tc}.
6706 @deftypefn {Target Hook} int TARGET_SCHED_SPECULATE_INSN (rtx @var{insn}, int @var{request}, rtx *@var{new_pat})
6707 This hook is called by the insn scheduler when @var{insn} has only
6708 speculative dependencies and therefore can be scheduled speculatively.
6709 The hook is used to check if the pattern of @var{insn} has a speculative
6710 version and, in case of successful check, to generate that speculative
6711 pattern. The hook should return 1, if the instruction has a speculative form,
6712 or @minus{}1, if it doesn't. @var{request} describes the type of requested
6713 speculation. If the return value equals 1 then @var{new_pat} is assigned
6714 the generated speculative pattern.
6717 @deftypefn {Target Hook} bool TARGET_SCHED_NEEDS_BLOCK_P (int @var{dep_status})
6718 This hook is called by the insn scheduler during generation of recovery code
6719 for @var{insn}. It should return @code{true}, if the corresponding check
6720 instruction should branch to recovery code, or @code{false} otherwise.
6723 @deftypefn {Target Hook} rtx TARGET_SCHED_GEN_SPEC_CHECK (rtx @var{insn}, rtx @var{label}, int @var{mutate_p})
6724 This hook is called by the insn scheduler to generate a pattern for recovery
6725 check instruction. If @var{mutate_p} is zero, then @var{insn} is a
6726 speculative instruction for which the check should be generated.
6727 @var{label} is either a label of a basic block, where recovery code should
6728 be emitted, or a null pointer, when requested check doesn't branch to
6729 recovery code (a simple check). If @var{mutate_p} is nonzero, then
6730 a pattern for a branchy check corresponding to a simple check denoted by
6731 @var{insn} should be generated. In this case @var{label} can't be null.
6734 @deftypefn {Target Hook} bool TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC (const_rtx @var{insn})
6735 This hook is used as a workaround for
6736 @samp{TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD} not being
6737 called on the first instruction of the ready list. The hook is used to
6738 discard speculative instructions that stand first in the ready list from
6739 being scheduled on the current cycle. If the hook returns @code{false},
6740 @var{insn} will not be chosen to be issued.
6741 For non-speculative instructions,
6742 the hook should always return @code{true}. For example, in the ia64 backend
6743 the hook is used to cancel data speculative insns when the ALAT table
6747 @deftypefn {Target Hook} void TARGET_SCHED_SET_SCHED_FLAGS (struct spec_info_def *@var{spec_info})
6748 This hook is used by the insn scheduler to find out what features should be
6750 The structure *@var{spec_info} should be filled in by the target.
6751 The structure describes speculation types that can be used in the scheduler.
6754 @deftypefn {Target Hook} int TARGET_SCHED_SMS_RES_MII (struct ddg *@var{g})
6755 This hook is called by the swing modulo scheduler to calculate a
6756 resource-based lower bound which is based on the resources available in
6757 the machine and the resources required by each instruction. The target
6758 backend can use @var{g} to calculate such bound. A very simple lower
6759 bound will be used in case this hook is not implemented: the total number
6760 of instructions divided by the issue rate.
6763 @deftypefn {Target Hook} bool TARGET_SCHED_DISPATCH (rtx @var{insn}, int @var{x})
6764 This hook is called by Haifa Scheduler. It returns true if dispatch scheduling
6765 is supported in hardware and the condition specified in the parameter is true.
6768 @deftypefn {Target Hook} void TARGET_SCHED_DISPATCH_DO (rtx @var{insn}, int @var{x})
6769 This hook is called by Haifa Scheduler. It performs the operation specified
6770 in its second parameter.
6774 @section Dividing the Output into Sections (Texts, Data, @dots{})
6775 @c the above section title is WAY too long. maybe cut the part between
6776 @c the (...)? --mew 10feb93
6778 An object file is divided into sections containing different types of
6779 data. In the most common case, there are three sections: the @dfn{text
6780 section}, which holds instructions and read-only data; the @dfn{data
6781 section}, which holds initialized writable data; and the @dfn{bss
6782 section}, which holds uninitialized data. Some systems have other kinds
6785 @file{varasm.c} provides several well-known sections, such as
6786 @code{text_section}, @code{data_section} and @code{bss_section}.
6787 The normal way of controlling a @code{@var{foo}_section} variable
6788 is to define the associated @code{@var{FOO}_SECTION_ASM_OP} macro,
6789 as described below. The macros are only read once, when @file{varasm.c}
6790 initializes itself, so their values must be run-time constants.
6791 They may however depend on command-line flags.
6793 @emph{Note:} Some run-time files, such @file{crtstuff.c}, also make
6794 use of the @code{@var{FOO}_SECTION_ASM_OP} macros, and expect them
6795 to be string literals.
6797 Some assemblers require a different string to be written every time a
6798 section is selected. If your assembler falls into this category, you
6799 should define the @code{TARGET_ASM_INIT_SECTIONS} hook and use
6800 @code{get_unnamed_section} to set up the sections.
6802 You must always create a @code{text_section}, either by defining
6803 @code{TEXT_SECTION_ASM_OP} or by initializing @code{text_section}
6804 in @code{TARGET_ASM_INIT_SECTIONS}. The same is true of
6805 @code{data_section} and @code{DATA_SECTION_ASM_OP}. If you do not
6806 create a distinct @code{readonly_data_section}, the default is to
6807 reuse @code{text_section}.
6809 All the other @file{varasm.c} sections are optional, and are null
6810 if the target does not provide them.
6812 @defmac TEXT_SECTION_ASM_OP
6813 A C expression whose value is a string, including spacing, containing the
6814 assembler operation that should precede instructions and read-only data.
6815 Normally @code{"\t.text"} is right.
6818 @defmac HOT_TEXT_SECTION_NAME
6819 If defined, a C string constant for the name of the section containing most
6820 frequently executed functions of the program. If not defined, GCC will provide
6821 a default definition if the target supports named sections.
6824 @defmac UNLIKELY_EXECUTED_TEXT_SECTION_NAME
6825 If defined, a C string constant for the name of the section containing unlikely
6826 executed functions in the program.
6829 @defmac DATA_SECTION_ASM_OP
6830 A C expression whose value is a string, including spacing, containing the
6831 assembler operation to identify the following data as writable initialized
6832 data. Normally @code{"\t.data"} is right.
6835 @defmac SDATA_SECTION_ASM_OP
6836 If defined, a C expression whose value is a string, including spacing,
6837 containing the assembler operation to identify the following data as
6838 initialized, writable small data.
6841 @defmac READONLY_DATA_SECTION_ASM_OP
6842 A C expression whose value is a string, including spacing, containing the
6843 assembler operation to identify the following data as read-only initialized
6847 @defmac BSS_SECTION_ASM_OP
6848 If defined, a C expression whose value is a string, including spacing,
6849 containing the assembler operation to identify the following data as
6850 uninitialized global data. If not defined, and neither
6851 @code{ASM_OUTPUT_BSS} nor @code{ASM_OUTPUT_ALIGNED_BSS} are defined,
6852 uninitialized global data will be output in the data section if
6853 @option{-fno-common} is passed, otherwise @code{ASM_OUTPUT_COMMON} will be
6857 @defmac SBSS_SECTION_ASM_OP
6858 If defined, a C expression whose value is a string, including spacing,
6859 containing the assembler operation to identify the following data as
6860 uninitialized, writable small data.
6863 @defmac TLS_COMMON_ASM_OP
6864 If defined, a C expression whose value is a string containing the
6865 assembler operation to identify the following data as thread-local
6866 common data. The default is @code{".tls_common"}.
6869 @defmac TLS_SECTION_ASM_FLAG
6870 If defined, a C expression whose value is a character constant
6871 containing the flag used to mark a section as a TLS section. The
6872 default is @code{'T'}.
6875 @defmac INIT_SECTION_ASM_OP
6876 If defined, a C expression whose value is a string, including spacing,
6877 containing the assembler operation to identify the following data as
6878 initialization code. If not defined, GCC will assume such a section does
6879 not exist. This section has no corresponding @code{init_section}
6880 variable; it is used entirely in runtime code.
6883 @defmac FINI_SECTION_ASM_OP
6884 If defined, a C expression whose value is a string, including spacing,
6885 containing the assembler operation to identify the following data as
6886 finalization code. If not defined, GCC will assume such a section does
6887 not exist. This section has no corresponding @code{fini_section}
6888 variable; it is used entirely in runtime code.
6891 @defmac INIT_ARRAY_SECTION_ASM_OP
6892 If defined, a C expression whose value is a string, including spacing,
6893 containing the assembler operation to identify the following data as
6894 part of the @code{.init_array} (or equivalent) section. If not
6895 defined, GCC will assume such a section does not exist. Do not define
6896 both this macro and @code{INIT_SECTION_ASM_OP}.
6899 @defmac FINI_ARRAY_SECTION_ASM_OP
6900 If defined, a C expression whose value is a string, including spacing,
6901 containing the assembler operation to identify the following data as
6902 part of the @code{.fini_array} (or equivalent) section. If not
6903 defined, GCC will assume such a section does not exist. Do not define
6904 both this macro and @code{FINI_SECTION_ASM_OP}.
6907 @defmac CRT_CALL_STATIC_FUNCTION (@var{section_op}, @var{function})
6908 If defined, an ASM statement that switches to a different section
6909 via @var{section_op}, calls @var{function}, and switches back to
6910 the text section. This is used in @file{crtstuff.c} if
6911 @code{INIT_SECTION_ASM_OP} or @code{FINI_SECTION_ASM_OP} to calls
6912 to initialization and finalization functions from the init and fini
6913 sections. By default, this macro uses a simple function call. Some
6914 ports need hand-crafted assembly code to avoid dependencies on
6915 registers initialized in the function prologue or to ensure that
6916 constant pools don't end up too far way in the text section.
6919 @defmac TARGET_LIBGCC_SDATA_SECTION
6920 If defined, a string which names the section into which small
6921 variables defined in crtstuff and libgcc should go. This is useful
6922 when the target has options for optimizing access to small data, and
6923 you want the crtstuff and libgcc routines to be conservative in what
6924 they expect of your application yet liberal in what your application
6925 expects. For example, for targets with a @code{.sdata} section (like
6926 MIPS), you could compile crtstuff with @code{-G 0} so that it doesn't
6927 require small data support from your application, but use this macro
6928 to put small data into @code{.sdata} so that your application can
6929 access these variables whether it uses small data or not.
6932 @defmac FORCE_CODE_SECTION_ALIGN
6933 If defined, an ASM statement that aligns a code section to some
6934 arbitrary boundary. This is used to force all fragments of the
6935 @code{.init} and @code{.fini} sections to have to same alignment
6936 and thus prevent the linker from having to add any padding.
6939 @defmac JUMP_TABLES_IN_TEXT_SECTION
6940 Define this macro to be an expression with a nonzero value if jump
6941 tables (for @code{tablejump} insns) should be output in the text
6942 section, along with the assembler instructions. Otherwise, the
6943 readonly data section is used.
6945 This macro is irrelevant if there is no separate readonly data section.
6948 @deftypefn {Target Hook} void TARGET_ASM_INIT_SECTIONS (void)
6949 Define this hook if you need to do something special to set up the
6950 @file{varasm.c} sections, or if your target has some special sections
6951 of its own that you need to create.
6953 GCC calls this hook after processing the command line, but before writing
6954 any assembly code, and before calling any of the section-returning hooks
6958 @deftypefn {Target Hook} int TARGET_ASM_RELOC_RW_MASK (void)
6959 Return a mask describing how relocations should be treated when
6960 selecting sections. Bit 1 should be set if global relocations
6961 should be placed in a read-write section; bit 0 should be set if
6962 local relocations should be placed in a read-write section.
6964 The default version of this function returns 3 when @option{-fpic}
6965 is in effect, and 0 otherwise. The hook is typically redefined
6966 when the target cannot support (some kinds of) dynamic relocations
6967 in read-only sections even in executables.
6970 @deftypefn {Target Hook} {section *} TARGET_ASM_SELECT_SECTION (tree @var{exp}, int @var{reloc}, unsigned HOST_WIDE_INT @var{align})
6971 Return the section into which @var{exp} should be placed. You can
6972 assume that @var{exp} is either a @code{VAR_DECL} node or a constant of
6973 some sort. @var{reloc} indicates whether the initial value of @var{exp}
6974 requires link-time relocations. Bit 0 is set when variable contains
6975 local relocations only, while bit 1 is set for global relocations.
6976 @var{align} is the constant alignment in bits.
6978 The default version of this function takes care of putting read-only
6979 variables in @code{readonly_data_section}.
6981 See also @var{USE_SELECT_SECTION_FOR_FUNCTIONS}.
6984 @defmac USE_SELECT_SECTION_FOR_FUNCTIONS
6985 Define this macro if you wish TARGET_ASM_SELECT_SECTION to be called
6986 for @code{FUNCTION_DECL}s as well as for variables and constants.
6988 In the case of a @code{FUNCTION_DECL}, @var{reloc} will be zero if the
6989 function has been determined to be likely to be called, and nonzero if
6990 it is unlikely to be called.
6993 @deftypefn {Target Hook} void TARGET_ASM_UNIQUE_SECTION (tree @var{decl}, int @var{reloc})
6994 Build up a unique section name, expressed as a @code{STRING_CST} node,
6995 and assign it to @samp{DECL_SECTION_NAME (@var{decl})}.
6996 As with @code{TARGET_ASM_SELECT_SECTION}, @var{reloc} indicates whether
6997 the initial value of @var{exp} requires link-time relocations.
6999 The default version of this function appends the symbol name to the
7000 ELF section name that would normally be used for the symbol. For
7001 example, the function @code{foo} would be placed in @code{.text.foo}.
7002 Whatever the actual target object format, this is often good enough.
7005 @deftypefn {Target Hook} {section *} TARGET_ASM_FUNCTION_RODATA_SECTION (tree @var{decl})
7006 Return the readonly data section associated with
7007 @samp{DECL_SECTION_NAME (@var{decl})}.
7008 The default version of this function selects @code{.gnu.linkonce.r.name} if
7009 the function's section is @code{.gnu.linkonce.t.name}, @code{.rodata.name}
7010 if function is in @code{.text.name}, and the normal readonly-data section
7014 @deftypefn {Target Hook} {section *} TARGET_ASM_SELECT_RTX_SECTION (enum machine_mode @var{mode}, rtx @var{x}, unsigned HOST_WIDE_INT @var{align})
7015 Return the section into which a constant @var{x}, of mode @var{mode},
7016 should be placed. You can assume that @var{x} is some kind of
7017 constant in RTL@. The argument @var{mode} is redundant except in the
7018 case of a @code{const_int} rtx. @var{align} is the constant alignment
7021 The default version of this function takes care of putting symbolic
7022 constants in @code{flag_pic} mode in @code{data_section} and everything
7023 else in @code{readonly_data_section}.
7026 @deftypefn {Target Hook} tree TARGET_MANGLE_DECL_ASSEMBLER_NAME (tree @var{decl}, tree @var{id})
7027 Define this hook if you need to postprocess the assembler name generated
7028 by target-independent code. The @var{id} provided to this hook will be
7029 the computed name (e.g., the macro @code{DECL_NAME} of the @var{decl} in C,
7030 or the mangled name of the @var{decl} in C++). The return value of the
7031 hook is an @code{IDENTIFIER_NODE} for the appropriate mangled name on
7032 your target system. The default implementation of this hook just
7033 returns the @var{id} provided.
7036 @deftypefn {Target Hook} void TARGET_ENCODE_SECTION_INFO (tree @var{decl}, rtx @var{rtl}, int @var{new_decl_p})
7037 Define this hook if references to a symbol or a constant must be
7038 treated differently depending on something about the variable or
7039 function named by the symbol (such as what section it is in).
7041 The hook is executed immediately after rtl has been created for
7042 @var{decl}, which may be a variable or function declaration or
7043 an entry in the constant pool. In either case, @var{rtl} is the
7044 rtl in question. Do @emph{not} use @code{DECL_RTL (@var{decl})}
7045 in this hook; that field may not have been initialized yet.
7047 In the case of a constant, it is safe to assume that the rtl is
7048 a @code{mem} whose address is a @code{symbol_ref}. Most decls
7049 will also have this form, but that is not guaranteed. Global
7050 register variables, for instance, will have a @code{reg} for their
7051 rtl. (Normally the right thing to do with such unusual rtl is
7054 The @var{new_decl_p} argument will be true if this is the first time
7055 that @code{TARGET_ENCODE_SECTION_INFO} has been invoked on this decl. It will
7056 be false for subsequent invocations, which will happen for duplicate
7057 declarations. Whether or not anything must be done for the duplicate
7058 declaration depends on whether the hook examines @code{DECL_ATTRIBUTES}.
7059 @var{new_decl_p} is always true when the hook is called for a constant.
7061 @cindex @code{SYMBOL_REF_FLAG}, in @code{TARGET_ENCODE_SECTION_INFO}
7062 The usual thing for this hook to do is to record flags in the
7063 @code{symbol_ref}, using @code{SYMBOL_REF_FLAG} or @code{SYMBOL_REF_FLAGS}.
7064 Historically, the name string was modified if it was necessary to
7065 encode more than one bit of information, but this practice is now
7066 discouraged; use @code{SYMBOL_REF_FLAGS}.
7068 The default definition of this hook, @code{default_encode_section_info}
7069 in @file{varasm.c}, sets a number of commonly-useful bits in
7070 @code{SYMBOL_REF_FLAGS}. Check whether the default does what you need
7071 before overriding it.
7074 @deftypefn {Target Hook} {const char *} TARGET_STRIP_NAME_ENCODING (const char *@var{name})
7075 Decode @var{name} and return the real name part, sans
7076 the characters that @code{TARGET_ENCODE_SECTION_INFO}
7080 @deftypefn {Target Hook} bool TARGET_IN_SMALL_DATA_P (const_tree @var{exp})
7081 Returns true if @var{exp} should be placed into a ``small data'' section.
7082 The default version of this hook always returns false.
7085 @deftypevr {Target Hook} bool TARGET_HAVE_SRODATA_SECTION
7086 Contains the value true if the target places read-only
7087 ``small data'' into a separate section. The default value is false.
7090 @deftypefn {Target Hook} bool TARGET_PROFILE_BEFORE_PROLOGUE (void)
7091 It returns true if target wants profile code emitted before prologue.
7093 The default version of this hook use the target macro
7094 @code{PROFILE_BEFORE_PROLOGUE}.
7097 @deftypefn {Target Hook} bool TARGET_BINDS_LOCAL_P (const_tree @var{exp})
7098 Returns true if @var{exp} names an object for which name resolution
7099 rules must resolve to the current ``module'' (dynamic shared library
7100 or executable image).
7102 The default version of this hook implements the name resolution rules
7103 for ELF, which has a looser model of global name binding than other
7104 currently supported object file formats.
7107 @deftypevr {Target Hook} bool TARGET_HAVE_TLS
7108 Contains the value true if the target supports thread-local storage.
7109 The default value is false.
7114 @section Position Independent Code
7115 @cindex position independent code
7118 This section describes macros that help implement generation of position
7119 independent code. Simply defining these macros is not enough to
7120 generate valid PIC; you must also add support to the hook
7121 @code{TARGET_LEGITIMATE_ADDRESS_P} and to the macro
7122 @code{PRINT_OPERAND_ADDRESS}, as well as @code{LEGITIMIZE_ADDRESS}. You
7123 must modify the definition of @samp{movsi} to do something appropriate
7124 when the source operand contains a symbolic address. You may also
7125 need to alter the handling of switch statements so that they use
7127 @c i rearranged the order of the macros above to try to force one of
7128 @c them to the next line, to eliminate an overfull hbox. --mew 10feb93
7130 @defmac PIC_OFFSET_TABLE_REGNUM
7131 The register number of the register used to address a table of static
7132 data addresses in memory. In some cases this register is defined by a
7133 processor's ``application binary interface'' (ABI)@. When this macro
7134 is defined, RTL is generated for this register once, as with the stack
7135 pointer and frame pointer registers. If this macro is not defined, it
7136 is up to the machine-dependent files to allocate such a register (if
7137 necessary). Note that this register must be fixed when in use (e.g.@:
7138 when @code{flag_pic} is true).
7141 @defmac PIC_OFFSET_TABLE_REG_CALL_CLOBBERED
7142 A C expression that is nonzero if the register defined by
7143 @code{PIC_OFFSET_TABLE_REGNUM} is clobbered by calls. If not defined,
7144 the default is zero. Do not define
7145 this macro if @code{PIC_OFFSET_TABLE_REGNUM} is not defined.
7148 @defmac LEGITIMATE_PIC_OPERAND_P (@var{x})
7149 A C expression that is nonzero if @var{x} is a legitimate immediate
7150 operand on the target machine when generating position independent code.
7151 You can assume that @var{x} satisfies @code{CONSTANT_P}, so you need not
7152 check this. You can also assume @var{flag_pic} is true, so you need not
7153 check it either. You need not define this macro if all constants
7154 (including @code{SYMBOL_REF}) can be immediate operands when generating
7155 position independent code.
7158 @node Assembler Format
7159 @section Defining the Output Assembler Language
7161 This section describes macros whose principal purpose is to describe how
7162 to write instructions in assembler language---rather than what the
7166 * File Framework:: Structural information for the assembler file.
7167 * Data Output:: Output of constants (numbers, strings, addresses).
7168 * Uninitialized Data:: Output of uninitialized variables.
7169 * Label Output:: Output and generation of labels.
7170 * Initialization:: General principles of initialization
7171 and termination routines.
7172 * Macros for Initialization::
7173 Specific macros that control the handling of
7174 initialization and termination routines.
7175 * Instruction Output:: Output of actual instructions.
7176 * Dispatch Tables:: Output of jump tables.
7177 * Exception Region Output:: Output of exception region code.
7178 * Alignment Output:: Pseudo ops for alignment and skipping data.
7181 @node File Framework
7182 @subsection The Overall Framework of an Assembler File
7183 @cindex assembler format
7184 @cindex output of assembler code
7186 @c prevent bad page break with this line
7187 This describes the overall framework of an assembly file.
7189 @findex default_file_start
7190 @deftypefn {Target Hook} void TARGET_ASM_FILE_START (void)
7191 Output to @code{asm_out_file} any text which the assembler expects to
7192 find at the beginning of a file. The default behavior is controlled
7193 by two flags, documented below. Unless your target's assembler is
7194 quite unusual, if you override the default, you should call
7195 @code{default_file_start} at some point in your target hook. This
7196 lets other target files rely on these variables.
7199 @deftypevr {Target Hook} bool TARGET_ASM_FILE_START_APP_OFF
7200 If this flag is true, the text of the macro @code{ASM_APP_OFF} will be
7201 printed as the very first line in the assembly file, unless
7202 @option{-fverbose-asm} is in effect. (If that macro has been defined
7203 to the empty string, this variable has no effect.) With the normal
7204 definition of @code{ASM_APP_OFF}, the effect is to notify the GNU
7205 assembler that it need not bother stripping comments or extra
7206 whitespace from its input. This allows it to work a bit faster.
7208 The default is false. You should not set it to true unless you have
7209 verified that your port does not generate any extra whitespace or
7210 comments that will cause GAS to issue errors in NO_APP mode.
7213 @deftypevr {Target Hook} bool TARGET_ASM_FILE_START_FILE_DIRECTIVE
7214 If this flag is true, @code{output_file_directive} will be called
7215 for the primary source file, immediately after printing
7216 @code{ASM_APP_OFF} (if that is enabled). Most ELF assemblers expect
7217 this to be done. The default is false.
7220 @deftypefn {Target Hook} void TARGET_ASM_FILE_END (void)
7221 Output to @code{asm_out_file} any text which the assembler expects
7222 to find at the end of a file. The default is to output nothing.
7225 @deftypefun void file_end_indicate_exec_stack ()
7226 Some systems use a common convention, the @samp{.note.GNU-stack}
7227 special section, to indicate whether or not an object file relies on
7228 the stack being executable. If your system uses this convention, you
7229 should define @code{TARGET_ASM_FILE_END} to this function. If you
7230 need to do other things in that hook, have your hook function call
7234 @deftypefn {Target Hook} void TARGET_ASM_LTO_START (void)
7235 Output to @code{asm_out_file} any text which the assembler expects
7236 to find at the start of an LTO section. The default is to output
7240 @deftypefn {Target Hook} void TARGET_ASM_LTO_END (void)
7241 Output to @code{asm_out_file} any text which the assembler expects
7242 to find at the end of an LTO section. The default is to output
7246 @deftypefn {Target Hook} void TARGET_ASM_CODE_END (void)
7247 Output to @code{asm_out_file} any text which is needed before emitting
7248 unwind info and debug info at the end of a file. Some targets emit
7249 here PIC setup thunks that cannot be emitted at the end of file,
7250 because they couldn't have unwind info then. The default is to output
7254 @defmac ASM_COMMENT_START
7255 A C string constant describing how to begin a comment in the target
7256 assembler language. The compiler assumes that the comment will end at
7257 the end of the line.
7261 A C string constant for text to be output before each @code{asm}
7262 statement or group of consecutive ones. Normally this is
7263 @code{"#APP"}, which is a comment that has no effect on most
7264 assemblers but tells the GNU assembler that it must check the lines
7265 that follow for all valid assembler constructs.
7269 A C string constant for text to be output after each @code{asm}
7270 statement or group of consecutive ones. Normally this is
7271 @code{"#NO_APP"}, which tells the GNU assembler to resume making the
7272 time-saving assumptions that are valid for ordinary compiler output.
7275 @defmac ASM_OUTPUT_SOURCE_FILENAME (@var{stream}, @var{name})
7276 A C statement to output COFF information or DWARF debugging information
7277 which indicates that filename @var{name} is the current source file to
7278 the stdio stream @var{stream}.
7280 This macro need not be defined if the standard form of output
7281 for the file format in use is appropriate.
7284 @deftypefn {Target Hook} void TARGET_ASM_OUTPUT_SOURCE_FILENAME (FILE *@var{file}, const char *@var{name})
7285 Output COFF information or DWARF debugging information which indicates that filename @var{name} is the current source file to the stdio stream @var{file}.
7287 This target hook need not be defined if the standard form of output for the file format in use is appropriate.
7290 @defmac OUTPUT_QUOTED_STRING (@var{stream}, @var{string})
7291 A C statement to output the string @var{string} to the stdio stream
7292 @var{stream}. If you do not call the function @code{output_quoted_string}
7293 in your config files, GCC will only call it to output filenames to
7294 the assembler source. So you can use it to canonicalize the format
7295 of the filename using this macro.
7298 @defmac ASM_OUTPUT_IDENT (@var{stream}, @var{string})
7299 A C statement to output something to the assembler file to handle a
7300 @samp{#ident} directive containing the text @var{string}. If this
7301 macro is not defined, nothing is output for a @samp{#ident} directive.
7304 @deftypefn {Target Hook} void TARGET_ASM_NAMED_SECTION (const char *@var{name}, unsigned int @var{flags}, tree @var{decl})
7305 Output assembly directives to switch to section @var{name}. The section
7306 should have attributes as specified by @var{flags}, which is a bit mask
7307 of the @code{SECTION_*} flags defined in @file{output.h}. If @var{decl}
7308 is non-NULL, it is the @code{VAR_DECL} or @code{FUNCTION_DECL} with which
7309 this section is associated.
7312 @deftypevr {Target Hook} bool TARGET_HAVE_NAMED_SECTIONS
7313 This flag is true if the target supports @code{TARGET_ASM_NAMED_SECTION}.
7316 @anchor{TARGET_HAVE_SWITCHABLE_BSS_SECTIONS}
7317 @deftypevr {Target Hook} bool TARGET_HAVE_SWITCHABLE_BSS_SECTIONS
7318 This flag is true if we can create zeroed data by switching to a BSS
7319 section and then using @code{ASM_OUTPUT_SKIP} to allocate the space.
7320 This is true on most ELF targets.
7323 @deftypefn {Target Hook} {unsigned int} TARGET_SECTION_TYPE_FLAGS (tree @var{decl}, const char *@var{name}, int @var{reloc})
7324 Choose a set of section attributes for use by @code{TARGET_ASM_NAMED_SECTION}
7325 based on a variable or function decl, a section name, and whether or not the
7326 declaration's initializer may contain runtime relocations. @var{decl} may be
7327 null, in which case read-write data should be assumed.
7329 The default version of this function handles choosing code vs data,
7330 read-only vs read-write data, and @code{flag_pic}. You should only
7331 need to override this if your target has special flags that might be
7332 set via @code{__attribute__}.
7335 @deftypefn {Target Hook} int TARGET_ASM_RECORD_GCC_SWITCHES (print_switch_type @var{type}, const char *@var{text})
7336 Provides the target with the ability to record the gcc command line
7337 switches that have been passed to the compiler, and options that are
7338 enabled. The @var{type} argument specifies what is being recorded.
7339 It can take the following values:
7342 @item SWITCH_TYPE_PASSED
7343 @var{text} is a command line switch that has been set by the user.
7345 @item SWITCH_TYPE_ENABLED
7346 @var{text} is an option which has been enabled. This might be as a
7347 direct result of a command line switch, or because it is enabled by
7348 default or because it has been enabled as a side effect of a different
7349 command line switch. For example, the @option{-O2} switch enables
7350 various different individual optimization passes.
7352 @item SWITCH_TYPE_DESCRIPTIVE
7353 @var{text} is either NULL or some descriptive text which should be
7354 ignored. If @var{text} is NULL then it is being used to warn the
7355 target hook that either recording is starting or ending. The first
7356 time @var{type} is SWITCH_TYPE_DESCRIPTIVE and @var{text} is NULL, the
7357 warning is for start up and the second time the warning is for
7358 wind down. This feature is to allow the target hook to make any
7359 necessary preparations before it starts to record switches and to
7360 perform any necessary tidying up after it has finished recording
7363 @item SWITCH_TYPE_LINE_START
7364 This option can be ignored by this target hook.
7366 @item SWITCH_TYPE_LINE_END
7367 This option can be ignored by this target hook.
7370 The hook's return value must be zero. Other return values may be
7371 supported in the future.
7373 By default this hook is set to NULL, but an example implementation is
7374 provided for ELF based targets. Called @var{elf_record_gcc_switches},
7375 it records the switches as ASCII text inside a new, string mergeable
7376 section in the assembler output file. The name of the new section is
7377 provided by the @code{TARGET_ASM_RECORD_GCC_SWITCHES_SECTION} target
7381 @deftypevr {Target Hook} {const char *} TARGET_ASM_RECORD_GCC_SWITCHES_SECTION
7382 This is the name of the section that will be created by the example
7383 ELF implementation of the @code{TARGET_ASM_RECORD_GCC_SWITCHES} target
7389 @subsection Output of Data
7392 @deftypevr {Target Hook} {const char *} TARGET_ASM_BYTE_OP
7393 @deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_HI_OP
7394 @deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_SI_OP
7395 @deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_DI_OP
7396 @deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_TI_OP
7397 @deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_HI_OP
7398 @deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_SI_OP
7399 @deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_DI_OP
7400 @deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_TI_OP
7401 These hooks specify assembly directives for creating certain kinds
7402 of integer object. The @code{TARGET_ASM_BYTE_OP} directive creates a
7403 byte-sized object, the @code{TARGET_ASM_ALIGNED_HI_OP} one creates an
7404 aligned two-byte object, and so on. Any of the hooks may be
7405 @code{NULL}, indicating that no suitable directive is available.
7407 The compiler will print these strings at the start of a new line,
7408 followed immediately by the object's initial value. In most cases,
7409 the string should contain a tab, a pseudo-op, and then another tab.
7412 @deftypefn {Target Hook} bool TARGET_ASM_INTEGER (rtx @var{x}, unsigned int @var{size}, int @var{aligned_p})
7413 The @code{assemble_integer} function uses this hook to output an
7414 integer object. @var{x} is the object's value, @var{size} is its size
7415 in bytes and @var{aligned_p} indicates whether it is aligned. The
7416 function should return @code{true} if it was able to output the
7417 object. If it returns false, @code{assemble_integer} will try to
7418 split the object into smaller parts.
7420 The default implementation of this hook will use the
7421 @code{TARGET_ASM_BYTE_OP} family of strings, returning @code{false}
7422 when the relevant string is @code{NULL}.
7425 @deftypefn {Target Hook} bool TARGET_ASM_OUTPUT_ADDR_CONST_EXTRA (FILE *@var{file}, rtx @var{x})
7426 A target hook to recognize @var{rtx} patterns that @code{output_addr_const}
7427 can't deal with, and output assembly code to @var{file} corresponding to
7428 the pattern @var{x}. This may be used to allow machine-dependent
7429 @code{UNSPEC}s to appear within constants.
7431 If target hook fails to recognize a pattern, it must return @code{false},
7432 so that a standard error message is printed. If it prints an error message
7433 itself, by calling, for example, @code{output_operand_lossage}, it may just
7437 @defmac OUTPUT_ADDR_CONST_EXTRA (@var{stream}, @var{x}, @var{fail})
7438 A C statement to recognize @var{rtx} patterns that
7439 @code{output_addr_const} can't deal with, and output assembly code to
7440 @var{stream} corresponding to the pattern @var{x}. This may be used to
7441 allow machine-dependent @code{UNSPEC}s to appear within constants.
7443 If @code{OUTPUT_ADDR_CONST_EXTRA} fails to recognize a pattern, it must
7444 @code{goto fail}, so that a standard error message is printed. If it
7445 prints an error message itself, by calling, for example,
7446 @code{output_operand_lossage}, it may just complete normally.
7449 @defmac ASM_OUTPUT_ASCII (@var{stream}, @var{ptr}, @var{len})
7450 A C statement to output to the stdio stream @var{stream} an assembler
7451 instruction to assemble a string constant containing the @var{len}
7452 bytes at @var{ptr}. @var{ptr} will be a C expression of type
7453 @code{char *} and @var{len} a C expression of type @code{int}.
7455 If the assembler has a @code{.ascii} pseudo-op as found in the
7456 Berkeley Unix assembler, do not define the macro
7457 @code{ASM_OUTPUT_ASCII}.
7460 @defmac ASM_OUTPUT_FDESC (@var{stream}, @var{decl}, @var{n})
7461 A C statement to output word @var{n} of a function descriptor for
7462 @var{decl}. This must be defined if @code{TARGET_VTABLE_USES_DESCRIPTORS}
7463 is defined, and is otherwise unused.
7466 @defmac CONSTANT_POOL_BEFORE_FUNCTION
7467 You may define this macro as a C expression. You should define the
7468 expression to have a nonzero value if GCC should output the constant
7469 pool for a function before the code for the function, or a zero value if
7470 GCC should output the constant pool after the function. If you do
7471 not define this macro, the usual case, GCC will output the constant
7472 pool before the function.
7475 @defmac ASM_OUTPUT_POOL_PROLOGUE (@var{file}, @var{funname}, @var{fundecl}, @var{size})
7476 A C statement to output assembler commands to define the start of the
7477 constant pool for a function. @var{funname} is a string giving
7478 the name of the function. Should the return type of the function
7479 be required, it can be obtained via @var{fundecl}. @var{size}
7480 is the size, in bytes, of the constant pool that will be written
7481 immediately after this call.
7483 If no constant-pool prefix is required, the usual case, this macro need
7487 @defmac ASM_OUTPUT_SPECIAL_POOL_ENTRY (@var{file}, @var{x}, @var{mode}, @var{align}, @var{labelno}, @var{jumpto})
7488 A C statement (with or without semicolon) to output a constant in the
7489 constant pool, if it needs special treatment. (This macro need not do
7490 anything for RTL expressions that can be output normally.)
7492 The argument @var{file} is the standard I/O stream to output the
7493 assembler code on. @var{x} is the RTL expression for the constant to
7494 output, and @var{mode} is the machine mode (in case @var{x} is a
7495 @samp{const_int}). @var{align} is the required alignment for the value
7496 @var{x}; you should output an assembler directive to force this much
7499 The argument @var{labelno} is a number to use in an internal label for
7500 the address of this pool entry. The definition of this macro is
7501 responsible for outputting the label definition at the proper place.
7502 Here is how to do this:
7505 @code{(*targetm.asm_out.internal_label)} (@var{file}, "LC", @var{labelno});
7508 When you output a pool entry specially, you should end with a
7509 @code{goto} to the label @var{jumpto}. This will prevent the same pool
7510 entry from being output a second time in the usual manner.
7512 You need not define this macro if it would do nothing.
7515 @defmac ASM_OUTPUT_POOL_EPILOGUE (@var{file} @var{funname} @var{fundecl} @var{size})
7516 A C statement to output assembler commands to at the end of the constant
7517 pool for a function. @var{funname} is a string giving the name of the
7518 function. Should the return type of the function be required, you can
7519 obtain it via @var{fundecl}. @var{size} is the size, in bytes, of the
7520 constant pool that GCC wrote immediately before this call.
7522 If no constant-pool epilogue is required, the usual case, you need not
7526 @defmac IS_ASM_LOGICAL_LINE_SEPARATOR (@var{C}, @var{STR})
7527 Define this macro as a C expression which is nonzero if @var{C} is
7528 used as a logical line separator by the assembler. @var{STR} points
7529 to the position in the string where @var{C} was found; this can be used if
7530 a line separator uses multiple characters.
7532 If you do not define this macro, the default is that only
7533 the character @samp{;} is treated as a logical line separator.
7536 @deftypevr {Target Hook} {const char *} TARGET_ASM_OPEN_PAREN
7537 @deftypevrx {Target Hook} {const char *} TARGET_ASM_CLOSE_PAREN
7538 These target hooks are C string constants, describing the syntax in the
7539 assembler for grouping arithmetic expressions. If not overridden, they
7540 default to normal parentheses, which is correct for most assemblers.
7543 These macros are provided by @file{real.h} for writing the definitions
7544 of @code{ASM_OUTPUT_DOUBLE} and the like:
7546 @defmac REAL_VALUE_TO_TARGET_SINGLE (@var{x}, @var{l})
7547 @defmacx REAL_VALUE_TO_TARGET_DOUBLE (@var{x}, @var{l})
7548 @defmacx REAL_VALUE_TO_TARGET_LONG_DOUBLE (@var{x}, @var{l})
7549 @defmacx REAL_VALUE_TO_TARGET_DECIMAL32 (@var{x}, @var{l})
7550 @defmacx REAL_VALUE_TO_TARGET_DECIMAL64 (@var{x}, @var{l})
7551 @defmacx REAL_VALUE_TO_TARGET_DECIMAL128 (@var{x}, @var{l})
7552 These translate @var{x}, of type @code{REAL_VALUE_TYPE}, to the
7553 target's floating point representation, and store its bit pattern in
7554 the variable @var{l}. For @code{REAL_VALUE_TO_TARGET_SINGLE} and
7555 @code{REAL_VALUE_TO_TARGET_DECIMAL32}, this variable should be a
7556 simple @code{long int}. For the others, it should be an array of
7557 @code{long int}. The number of elements in this array is determined
7558 by the size of the desired target floating point data type: 32 bits of
7559 it go in each @code{long int} array element. Each array element holds
7560 32 bits of the result, even if @code{long int} is wider than 32 bits
7561 on the host machine.
7563 The array element values are designed so that you can print them out
7564 using @code{fprintf} in the order they should appear in the target
7568 @node Uninitialized Data
7569 @subsection Output of Uninitialized Variables
7571 Each of the macros in this section is used to do the whole job of
7572 outputting a single uninitialized variable.
7574 @defmac ASM_OUTPUT_COMMON (@var{stream}, @var{name}, @var{size}, @var{rounded})
7575 A C statement (sans semicolon) to output to the stdio stream
7576 @var{stream} the assembler definition of a common-label named
7577 @var{name} whose size is @var{size} bytes. The variable @var{rounded}
7578 is the size rounded up to whatever alignment the caller wants. It is
7579 possible that @var{size} may be zero, for instance if a struct with no
7580 other member than a zero-length array is defined. In this case, the
7581 backend must output a symbol definition that allocates at least one
7582 byte, both so that the address of the resulting object does not compare
7583 equal to any other, and because some object formats cannot even express
7584 the concept of a zero-sized common symbol, as that is how they represent
7585 an ordinary undefined external.
7587 Use the expression @code{assemble_name (@var{stream}, @var{name})} to
7588 output the name itself; before and after that, output the additional
7589 assembler syntax for defining the name, and a newline.
7591 This macro controls how the assembler definitions of uninitialized
7592 common global variables are output.
7595 @defmac ASM_OUTPUT_ALIGNED_COMMON (@var{stream}, @var{name}, @var{size}, @var{alignment})
7596 Like @code{ASM_OUTPUT_COMMON} except takes the required alignment as a
7597 separate, explicit argument. If you define this macro, it is used in
7598 place of @code{ASM_OUTPUT_COMMON}, and gives you more flexibility in
7599 handling the required alignment of the variable. The alignment is specified
7600 as the number of bits.
7603 @defmac ASM_OUTPUT_ALIGNED_DECL_COMMON (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{alignment})
7604 Like @code{ASM_OUTPUT_ALIGNED_COMMON} except that @var{decl} of the
7605 variable to be output, if there is one, or @code{NULL_TREE} if there
7606 is no corresponding variable. If you define this macro, GCC will use it
7607 in place of both @code{ASM_OUTPUT_COMMON} and
7608 @code{ASM_OUTPUT_ALIGNED_COMMON}. Define this macro when you need to see
7609 the variable's decl in order to chose what to output.
7612 @defmac ASM_OUTPUT_BSS (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{rounded})
7613 A C statement (sans semicolon) to output to the stdio stream
7614 @var{stream} the assembler definition of uninitialized global @var{decl} named
7615 @var{name} whose size is @var{size} bytes. The variable @var{rounded}
7616 is the size rounded up to whatever alignment the caller wants.
7618 Try to use function @code{asm_output_bss} defined in @file{varasm.c} when
7619 defining this macro. If unable, use the expression
7620 @code{assemble_name (@var{stream}, @var{name})} to output the name itself;
7621 before and after that, output the additional assembler syntax for defining
7622 the name, and a newline.
7624 There are two ways of handling global BSS@. One is to define either
7625 this macro or its aligned counterpart, @code{ASM_OUTPUT_ALIGNED_BSS}.
7626 The other is to have @code{TARGET_ASM_SELECT_SECTION} return a
7627 switchable BSS section (@pxref{TARGET_HAVE_SWITCHABLE_BSS_SECTIONS}).
7628 You do not need to do both.
7630 Some languages do not have @code{common} data, and require a
7631 non-common form of global BSS in order to handle uninitialized globals
7632 efficiently. C++ is one example of this. However, if the target does
7633 not support global BSS, the front end may choose to make globals
7634 common in order to save space in the object file.
7637 @defmac ASM_OUTPUT_ALIGNED_BSS (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{alignment})
7638 Like @code{ASM_OUTPUT_BSS} except takes the required alignment as a
7639 separate, explicit argument. If you define this macro, it is used in
7640 place of @code{ASM_OUTPUT_BSS}, and gives you more flexibility in
7641 handling the required alignment of the variable. The alignment is specified
7642 as the number of bits.
7644 Try to use function @code{asm_output_aligned_bss} defined in file
7645 @file{varasm.c} when defining this macro.
7648 @defmac ASM_OUTPUT_LOCAL (@var{stream}, @var{name}, @var{size}, @var{rounded})
7649 A C statement (sans semicolon) to output to the stdio stream
7650 @var{stream} the assembler definition of a local-common-label named
7651 @var{name} whose size is @var{size} bytes. The variable @var{rounded}
7652 is the size rounded up to whatever alignment the caller wants.
7654 Use the expression @code{assemble_name (@var{stream}, @var{name})} to
7655 output the name itself; before and after that, output the additional
7656 assembler syntax for defining the name, and a newline.
7658 This macro controls how the assembler definitions of uninitialized
7659 static variables are output.
7662 @defmac ASM_OUTPUT_ALIGNED_LOCAL (@var{stream}, @var{name}, @var{size}, @var{alignment})
7663 Like @code{ASM_OUTPUT_LOCAL} except takes the required alignment as a
7664 separate, explicit argument. If you define this macro, it is used in
7665 place of @code{ASM_OUTPUT_LOCAL}, and gives you more flexibility in
7666 handling the required alignment of the variable. The alignment is specified
7667 as the number of bits.
7670 @defmac ASM_OUTPUT_ALIGNED_DECL_LOCAL (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{alignment})
7671 Like @code{ASM_OUTPUT_ALIGNED_DECL} except that @var{decl} of the
7672 variable to be output, if there is one, or @code{NULL_TREE} if there
7673 is no corresponding variable. If you define this macro, GCC will use it
7674 in place of both @code{ASM_OUTPUT_DECL} and
7675 @code{ASM_OUTPUT_ALIGNED_DECL}. Define this macro when you need to see
7676 the variable's decl in order to chose what to output.
7680 @subsection Output and Generation of Labels
7682 @c prevent bad page break with this line
7683 This is about outputting labels.
7685 @findex assemble_name
7686 @defmac ASM_OUTPUT_LABEL (@var{stream}, @var{name})
7687 A C statement (sans semicolon) to output to the stdio stream
7688 @var{stream} the assembler definition of a label named @var{name}.
7689 Use the expression @code{assemble_name (@var{stream}, @var{name})} to
7690 output the name itself; before and after that, output the additional
7691 assembler syntax for defining the name, and a newline. A default
7692 definition of this macro is provided which is correct for most systems.
7695 @defmac ASM_OUTPUT_FUNCTION_LABEL (@var{stream}, @var{name}, @var{decl})
7696 A C statement (sans semicolon) to output to the stdio stream
7697 @var{stream} the assembler definition of a label named @var{name} of
7699 Use the expression @code{assemble_name (@var{stream}, @var{name})} to
7700 output the name itself; before and after that, output the additional
7701 assembler syntax for defining the name, and a newline. A default
7702 definition of this macro is provided which is correct for most systems.
7704 If this macro is not defined, then the function name is defined in the
7705 usual manner as a label (by means of @code{ASM_OUTPUT_LABEL}).
7708 @findex assemble_name_raw
7709 @defmac ASM_OUTPUT_INTERNAL_LABEL (@var{stream}, @var{name})
7710 Identical to @code{ASM_OUTPUT_LABEL}, except that @var{name} is known
7711 to refer to a compiler-generated label. The default definition uses
7712 @code{assemble_name_raw}, which is like @code{assemble_name} except
7713 that it is more efficient.
7717 A C string containing the appropriate assembler directive to specify the
7718 size of a symbol, without any arguments. On systems that use ELF, the
7719 default (in @file{config/elfos.h}) is @samp{"\t.size\t"}; on other
7720 systems, the default is not to define this macro.
7722 Define this macro only if it is correct to use the default definitions
7723 of @code{ASM_OUTPUT_SIZE_DIRECTIVE} and @code{ASM_OUTPUT_MEASURED_SIZE}
7724 for your system. If you need your own custom definitions of those
7725 macros, or if you do not need explicit symbol sizes at all, do not
7729 @defmac ASM_OUTPUT_SIZE_DIRECTIVE (@var{stream}, @var{name}, @var{size})
7730 A C statement (sans semicolon) to output to the stdio stream
7731 @var{stream} a directive telling the assembler that the size of the
7732 symbol @var{name} is @var{size}. @var{size} is a @code{HOST_WIDE_INT}.
7733 If you define @code{SIZE_ASM_OP}, a default definition of this macro is
7737 @defmac ASM_OUTPUT_MEASURED_SIZE (@var{stream}, @var{name})
7738 A C statement (sans semicolon) to output to the stdio stream
7739 @var{stream} a directive telling the assembler to calculate the size of
7740 the symbol @var{name} by subtracting its address from the current
7743 If you define @code{SIZE_ASM_OP}, a default definition of this macro is
7744 provided. The default assumes that the assembler recognizes a special
7745 @samp{.} symbol as referring to the current address, and can calculate
7746 the difference between this and another symbol. If your assembler does
7747 not recognize @samp{.} or cannot do calculations with it, you will need
7748 to redefine @code{ASM_OUTPUT_MEASURED_SIZE} to use some other technique.
7752 A C string containing the appropriate assembler directive to specify the
7753 type of a symbol, without any arguments. On systems that use ELF, the
7754 default (in @file{config/elfos.h}) is @samp{"\t.type\t"}; on other
7755 systems, the default is not to define this macro.
7757 Define this macro only if it is correct to use the default definition of
7758 @code{ASM_OUTPUT_TYPE_DIRECTIVE} for your system. If you need your own
7759 custom definition of this macro, or if you do not need explicit symbol
7760 types at all, do not define this macro.
7763 @defmac TYPE_OPERAND_FMT
7764 A C string which specifies (using @code{printf} syntax) the format of
7765 the second operand to @code{TYPE_ASM_OP}. On systems that use ELF, the
7766 default (in @file{config/elfos.h}) is @samp{"@@%s"}; on other systems,
7767 the default is not to define this macro.
7769 Define this macro only if it is correct to use the default definition of
7770 @code{ASM_OUTPUT_TYPE_DIRECTIVE} for your system. If you need your own
7771 custom definition of this macro, or if you do not need explicit symbol
7772 types at all, do not define this macro.
7775 @defmac ASM_OUTPUT_TYPE_DIRECTIVE (@var{stream}, @var{type})
7776 A C statement (sans semicolon) to output to the stdio stream
7777 @var{stream} a directive telling the assembler that the type of the
7778 symbol @var{name} is @var{type}. @var{type} is a C string; currently,
7779 that string is always either @samp{"function"} or @samp{"object"}, but
7780 you should not count on this.
7782 If you define @code{TYPE_ASM_OP} and @code{TYPE_OPERAND_FMT}, a default
7783 definition of this macro is provided.
7786 @defmac ASM_DECLARE_FUNCTION_NAME (@var{stream}, @var{name}, @var{decl})
7787 A C statement (sans semicolon) to output to the stdio stream
7788 @var{stream} any text necessary for declaring the name @var{name} of a
7789 function which is being defined. This macro is responsible for
7790 outputting the label definition (perhaps using
7791 @code{ASM_OUTPUT_FUNCTION_LABEL}). The argument @var{decl} is the
7792 @code{FUNCTION_DECL} tree node representing the function.
7794 If this macro is not defined, then the function name is defined in the
7795 usual manner as a label (by means of @code{ASM_OUTPUT_FUNCTION_LABEL}).
7797 You may wish to use @code{ASM_OUTPUT_TYPE_DIRECTIVE} in the definition
7801 @defmac ASM_DECLARE_FUNCTION_SIZE (@var{stream}, @var{name}, @var{decl})
7802 A C statement (sans semicolon) to output to the stdio stream
7803 @var{stream} any text necessary for declaring the size of a function
7804 which is being defined. The argument @var{name} is the name of the
7805 function. The argument @var{decl} is the @code{FUNCTION_DECL} tree node
7806 representing the function.
7808 If this macro is not defined, then the function size is not defined.
7810 You may wish to use @code{ASM_OUTPUT_MEASURED_SIZE} in the definition
7814 @defmac ASM_DECLARE_OBJECT_NAME (@var{stream}, @var{name}, @var{decl})
7815 A C statement (sans semicolon) to output to the stdio stream
7816 @var{stream} any text necessary for declaring the name @var{name} of an
7817 initialized variable which is being defined. This macro must output the
7818 label definition (perhaps using @code{ASM_OUTPUT_LABEL}). The argument
7819 @var{decl} is the @code{VAR_DECL} tree node representing the variable.
7821 If this macro is not defined, then the variable name is defined in the
7822 usual manner as a label (by means of @code{ASM_OUTPUT_LABEL}).
7824 You may wish to use @code{ASM_OUTPUT_TYPE_DIRECTIVE} and/or
7825 @code{ASM_OUTPUT_SIZE_DIRECTIVE} in the definition of this macro.
7828 @deftypefn {Target Hook} void TARGET_ASM_DECLARE_CONSTANT_NAME (FILE *@var{file}, const char *@var{name}, const_tree @var{expr}, HOST_WIDE_INT @var{size})
7829 A target hook to output to the stdio stream @var{file} any text necessary
7830 for declaring the name @var{name} of a constant which is being defined. This
7831 target hook is responsible for outputting the label definition (perhaps using
7832 @code{assemble_label}). The argument @var{exp} is the value of the constant,
7833 and @var{size} is the size of the constant in bytes. The @var{name}
7834 will be an internal label.
7836 The default version of this target hook, define the @var{name} in the
7837 usual manner as a label (by means of @code{assemble_label}).
7839 You may wish to use @code{ASM_OUTPUT_TYPE_DIRECTIVE} in this target hook.
7842 @defmac ASM_DECLARE_REGISTER_GLOBAL (@var{stream}, @var{decl}, @var{regno}, @var{name})
7843 A C statement (sans semicolon) to output to the stdio stream
7844 @var{stream} any text necessary for claiming a register @var{regno}
7845 for a global variable @var{decl} with name @var{name}.
7847 If you don't define this macro, that is equivalent to defining it to do
7851 @defmac ASM_FINISH_DECLARE_OBJECT (@var{stream}, @var{decl}, @var{toplevel}, @var{atend})
7852 A C statement (sans semicolon) to finish up declaring a variable name
7853 once the compiler has processed its initializer fully and thus has had a
7854 chance to determine the size of an array when controlled by an
7855 initializer. This is used on systems where it's necessary to declare
7856 something about the size of the object.
7858 If you don't define this macro, that is equivalent to defining it to do
7861 You may wish to use @code{ASM_OUTPUT_SIZE_DIRECTIVE} and/or
7862 @code{ASM_OUTPUT_MEASURED_SIZE} in the definition of this macro.
7865 @deftypefn {Target Hook} void TARGET_ASM_GLOBALIZE_LABEL (FILE *@var{stream}, const char *@var{name})
7866 This target hook is a function to output to the stdio stream
7867 @var{stream} some commands that will make the label @var{name} global;
7868 that is, available for reference from other files.
7870 The default implementation relies on a proper definition of
7871 @code{GLOBAL_ASM_OP}.
7874 @deftypefn {Target Hook} void TARGET_ASM_GLOBALIZE_DECL_NAME (FILE *@var{stream}, tree @var{decl})
7875 This target hook is a function to output to the stdio stream
7876 @var{stream} some commands that will make the name associated with @var{decl}
7877 global; that is, available for reference from other files.
7879 The default implementation uses the TARGET_ASM_GLOBALIZE_LABEL target hook.
7882 @defmac ASM_WEAKEN_LABEL (@var{stream}, @var{name})
7883 A C statement (sans semicolon) to output to the stdio stream
7884 @var{stream} some commands that will make the label @var{name} weak;
7885 that is, available for reference from other files but only used if
7886 no other definition is available. Use the expression
7887 @code{assemble_name (@var{stream}, @var{name})} to output the name
7888 itself; before and after that, output the additional assembler syntax
7889 for making that name weak, and a newline.
7891 If you don't define this macro or @code{ASM_WEAKEN_DECL}, GCC will not
7892 support weak symbols and you should not define the @code{SUPPORTS_WEAK}
7896 @defmac ASM_WEAKEN_DECL (@var{stream}, @var{decl}, @var{name}, @var{value})
7897 Combines (and replaces) the function of @code{ASM_WEAKEN_LABEL} and
7898 @code{ASM_OUTPUT_WEAK_ALIAS}, allowing access to the associated function
7899 or variable decl. If @var{value} is not @code{NULL}, this C statement
7900 should output to the stdio stream @var{stream} assembler code which
7901 defines (equates) the weak symbol @var{name} to have the value
7902 @var{value}. If @var{value} is @code{NULL}, it should output commands
7903 to make @var{name} weak.
7906 @defmac ASM_OUTPUT_WEAKREF (@var{stream}, @var{decl}, @var{name}, @var{value})
7907 Outputs a directive that enables @var{name} to be used to refer to
7908 symbol @var{value} with weak-symbol semantics. @code{decl} is the
7909 declaration of @code{name}.
7912 @defmac SUPPORTS_WEAK
7913 A C expression which evaluates to true if the target supports weak symbols.
7915 If you don't define this macro, @file{defaults.h} provides a default
7916 definition. If either @code{ASM_WEAKEN_LABEL} or @code{ASM_WEAKEN_DECL}
7917 is defined, the default definition is @samp{1}; otherwise, it is
7918 @samp{0}. Define this macro if you want to control weak symbol support
7919 with a compiler flag such as @option{-melf}.
7922 @defmac MAKE_DECL_ONE_ONLY (@var{decl})
7923 A C statement (sans semicolon) to mark @var{decl} to be emitted as a
7924 public symbol such that extra copies in multiple translation units will
7925 be discarded by the linker. Define this macro if your object file
7926 format provides support for this concept, such as the @samp{COMDAT}
7927 section flags in the Microsoft Windows PE/COFF format, and this support
7928 requires changes to @var{decl}, such as putting it in a separate section.
7931 @defmac SUPPORTS_ONE_ONLY
7932 A C expression which evaluates to true if the target supports one-only
7935 If you don't define this macro, @file{varasm.c} provides a default
7936 definition. If @code{MAKE_DECL_ONE_ONLY} is defined, the default
7937 definition is @samp{1}; otherwise, it is @samp{0}. Define this macro if
7938 you want to control one-only symbol support with a compiler flag, or if
7939 setting the @code{DECL_ONE_ONLY} flag is enough to mark a declaration to
7940 be emitted as one-only.
7943 @deftypefn {Target Hook} void TARGET_ASM_ASSEMBLE_VISIBILITY (tree @var{decl}, int @var{visibility})
7944 This target hook is a function to output to @var{asm_out_file} some
7945 commands that will make the symbol(s) associated with @var{decl} have
7946 hidden, protected or internal visibility as specified by @var{visibility}.
7949 @defmac TARGET_WEAK_NOT_IN_ARCHIVE_TOC
7950 A C expression that evaluates to true if the target's linker expects
7951 that weak symbols do not appear in a static archive's table of contents.
7952 The default is @code{0}.
7954 Leaving weak symbols out of an archive's table of contents means that,
7955 if a symbol will only have a definition in one translation unit and
7956 will have undefined references from other translation units, that
7957 symbol should not be weak. Defining this macro to be nonzero will
7958 thus have the effect that certain symbols that would normally be weak
7959 (explicit template instantiations, and vtables for polymorphic classes
7960 with noninline key methods) will instead be nonweak.
7962 The C++ ABI requires this macro to be zero. Define this macro for
7963 targets where full C++ ABI compliance is impossible and where linker
7964 restrictions require weak symbols to be left out of a static archive's
7968 @defmac ASM_OUTPUT_EXTERNAL (@var{stream}, @var{decl}, @var{name})
7969 A C statement (sans semicolon) to output to the stdio stream
7970 @var{stream} any text necessary for declaring the name of an external
7971 symbol named @var{name} which is referenced in this compilation but
7972 not defined. The value of @var{decl} is the tree node for the
7975 This macro need not be defined if it does not need to output anything.
7976 The GNU assembler and most Unix assemblers don't require anything.
7979 @deftypefn {Target Hook} void TARGET_ASM_EXTERNAL_LIBCALL (rtx @var{symref})
7980 This target hook is a function to output to @var{asm_out_file} an assembler
7981 pseudo-op to declare a library function name external. The name of the
7982 library function is given by @var{symref}, which is a @code{symbol_ref}.
7985 @deftypefn {Target Hook} void TARGET_ASM_MARK_DECL_PRESERVED (const char *@var{symbol})
7986 This target hook is a function to output to @var{asm_out_file} an assembler
7987 directive to annotate @var{symbol} as used. The Darwin target uses the
7988 .no_dead_code_strip directive.
7991 @defmac ASM_OUTPUT_LABELREF (@var{stream}, @var{name})
7992 A C statement (sans semicolon) to output to the stdio stream
7993 @var{stream} a reference in assembler syntax to a label named
7994 @var{name}. This should add @samp{_} to the front of the name, if that
7995 is customary on your operating system, as it is in most Berkeley Unix
7996 systems. This macro is used in @code{assemble_name}.
7999 @defmac ASM_OUTPUT_SYMBOL_REF (@var{stream}, @var{sym})
8000 A C statement (sans semicolon) to output a reference to
8001 @code{SYMBOL_REF} @var{sym}. If not defined, @code{assemble_name}
8002 will be used to output the name of the symbol. This macro may be used
8003 to modify the way a symbol is referenced depending on information
8004 encoded by @code{TARGET_ENCODE_SECTION_INFO}.
8007 @defmac ASM_OUTPUT_LABEL_REF (@var{stream}, @var{buf})
8008 A C statement (sans semicolon) to output a reference to @var{buf}, the
8009 result of @code{ASM_GENERATE_INTERNAL_LABEL}. If not defined,
8010 @code{assemble_name} will be used to output the name of the symbol.
8011 This macro is not used by @code{output_asm_label}, or the @code{%l}
8012 specifier that calls it; the intention is that this macro should be set
8013 when it is necessary to output a label differently when its address is
8017 @deftypefn {Target Hook} void TARGET_ASM_INTERNAL_LABEL (FILE *@var{stream}, const char *@var{prefix}, unsigned long @var{labelno})
8018 A function to output to the stdio stream @var{stream} a label whose
8019 name is made from the string @var{prefix} and the number @var{labelno}.
8021 It is absolutely essential that these labels be distinct from the labels
8022 used for user-level functions and variables. Otherwise, certain programs
8023 will have name conflicts with internal labels.
8025 It is desirable to exclude internal labels from the symbol table of the
8026 object file. Most assemblers have a naming convention for labels that
8027 should be excluded; on many systems, the letter @samp{L} at the
8028 beginning of a label has this effect. You should find out what
8029 convention your system uses, and follow it.
8031 The default version of this function utilizes @code{ASM_GENERATE_INTERNAL_LABEL}.
8034 @defmac ASM_OUTPUT_DEBUG_LABEL (@var{stream}, @var{prefix}, @var{num})
8035 A C statement to output to the stdio stream @var{stream} a debug info
8036 label whose name is made from the string @var{prefix} and the number
8037 @var{num}. This is useful for VLIW targets, where debug info labels
8038 may need to be treated differently than branch target labels. On some
8039 systems, branch target labels must be at the beginning of instruction
8040 bundles, but debug info labels can occur in the middle of instruction
8043 If this macro is not defined, then @code{(*targetm.asm_out.internal_label)} will be
8047 @defmac ASM_GENERATE_INTERNAL_LABEL (@var{string}, @var{prefix}, @var{num})
8048 A C statement to store into the string @var{string} a label whose name
8049 is made from the string @var{prefix} and the number @var{num}.
8051 This string, when output subsequently by @code{assemble_name}, should
8052 produce the output that @code{(*targetm.asm_out.internal_label)} would produce
8053 with the same @var{prefix} and @var{num}.
8055 If the string begins with @samp{*}, then @code{assemble_name} will
8056 output the rest of the string unchanged. It is often convenient for
8057 @code{ASM_GENERATE_INTERNAL_LABEL} to use @samp{*} in this way. If the
8058 string doesn't start with @samp{*}, then @code{ASM_OUTPUT_LABELREF} gets
8059 to output the string, and may change it. (Of course,
8060 @code{ASM_OUTPUT_LABELREF} is also part of your machine description, so
8061 you should know what it does on your machine.)
8064 @defmac ASM_FORMAT_PRIVATE_NAME (@var{outvar}, @var{name}, @var{number})
8065 A C expression to assign to @var{outvar} (which is a variable of type
8066 @code{char *}) a newly allocated string made from the string
8067 @var{name} and the number @var{number}, with some suitable punctuation
8068 added. Use @code{alloca} to get space for the string.
8070 The string will be used as an argument to @code{ASM_OUTPUT_LABELREF} to
8071 produce an assembler label for an internal static variable whose name is
8072 @var{name}. Therefore, the string must be such as to result in valid
8073 assembler code. The argument @var{number} is different each time this
8074 macro is executed; it prevents conflicts between similarly-named
8075 internal static variables in different scopes.
8077 Ideally this string should not be a valid C identifier, to prevent any
8078 conflict with the user's own symbols. Most assemblers allow periods
8079 or percent signs in assembler symbols; putting at least one of these
8080 between the name and the number will suffice.
8082 If this macro is not defined, a default definition will be provided
8083 which is correct for most systems.
8086 @defmac ASM_OUTPUT_DEF (@var{stream}, @var{name}, @var{value})
8087 A C statement to output to the stdio stream @var{stream} assembler code
8088 which defines (equates) the symbol @var{name} to have the value @var{value}.
8091 If @code{SET_ASM_OP} is defined, a default definition is provided which is
8092 correct for most systems.
8095 @defmac ASM_OUTPUT_DEF_FROM_DECLS (@var{stream}, @var{decl_of_name}, @var{decl_of_value})
8096 A C statement to output to the stdio stream @var{stream} assembler code
8097 which defines (equates) the symbol whose tree node is @var{decl_of_name}
8098 to have the value of the tree node @var{decl_of_value}. This macro will
8099 be used in preference to @samp{ASM_OUTPUT_DEF} if it is defined and if
8100 the tree nodes are available.
8103 If @code{SET_ASM_OP} is defined, a default definition is provided which is
8104 correct for most systems.
8107 @defmac TARGET_DEFERRED_OUTPUT_DEFS (@var{decl_of_name}, @var{decl_of_value})
8108 A C statement that evaluates to true if the assembler code which defines
8109 (equates) the symbol whose tree node is @var{decl_of_name} to have the value
8110 of the tree node @var{decl_of_value} should be emitted near the end of the
8111 current compilation unit. The default is to not defer output of defines.
8112 This macro affects defines output by @samp{ASM_OUTPUT_DEF} and
8113 @samp{ASM_OUTPUT_DEF_FROM_DECLS}.
8116 @defmac ASM_OUTPUT_WEAK_ALIAS (@var{stream}, @var{name}, @var{value})
8117 A C statement to output to the stdio stream @var{stream} assembler code
8118 which defines (equates) the weak symbol @var{name} to have the value
8119 @var{value}. If @var{value} is @code{NULL}, it defines @var{name} as
8120 an undefined weak symbol.
8122 Define this macro if the target only supports weak aliases; define
8123 @code{ASM_OUTPUT_DEF} instead if possible.
8126 @defmac OBJC_GEN_METHOD_LABEL (@var{buf}, @var{is_inst}, @var{class_name}, @var{cat_name}, @var{sel_name})
8127 Define this macro to override the default assembler names used for
8128 Objective-C methods.
8130 The default name is a unique method number followed by the name of the
8131 class (e.g.@: @samp{_1_Foo}). For methods in categories, the name of
8132 the category is also included in the assembler name (e.g.@:
8135 These names are safe on most systems, but make debugging difficult since
8136 the method's selector is not present in the name. Therefore, particular
8137 systems define other ways of computing names.
8139 @var{buf} is an expression of type @code{char *} which gives you a
8140 buffer in which to store the name; its length is as long as
8141 @var{class_name}, @var{cat_name} and @var{sel_name} put together, plus
8142 50 characters extra.
8144 The argument @var{is_inst} specifies whether the method is an instance
8145 method or a class method; @var{class_name} is the name of the class;
8146 @var{cat_name} is the name of the category (or @code{NULL} if the method is not
8147 in a category); and @var{sel_name} is the name of the selector.
8149 On systems where the assembler can handle quoted names, you can use this
8150 macro to provide more human-readable names.
8153 @defmac ASM_DECLARE_CLASS_REFERENCE (@var{stream}, @var{name})
8154 A C statement (sans semicolon) to output to the stdio stream
8155 @var{stream} commands to declare that the label @var{name} is an
8156 Objective-C class reference. This is only needed for targets whose
8157 linkers have special support for NeXT-style runtimes.
8160 @defmac ASM_DECLARE_UNRESOLVED_REFERENCE (@var{stream}, @var{name})
8161 A C statement (sans semicolon) to output to the stdio stream
8162 @var{stream} commands to declare that the label @var{name} is an
8163 unresolved Objective-C class reference. This is only needed for targets
8164 whose linkers have special support for NeXT-style runtimes.
8167 @node Initialization
8168 @subsection How Initialization Functions Are Handled
8169 @cindex initialization routines
8170 @cindex termination routines
8171 @cindex constructors, output of
8172 @cindex destructors, output of
8174 The compiled code for certain languages includes @dfn{constructors}
8175 (also called @dfn{initialization routines})---functions to initialize
8176 data in the program when the program is started. These functions need
8177 to be called before the program is ``started''---that is to say, before
8178 @code{main} is called.
8180 Compiling some languages generates @dfn{destructors} (also called
8181 @dfn{termination routines}) that should be called when the program
8184 To make the initialization and termination functions work, the compiler
8185 must output something in the assembler code to cause those functions to
8186 be called at the appropriate time. When you port the compiler to a new
8187 system, you need to specify how to do this.
8189 There are two major ways that GCC currently supports the execution of
8190 initialization and termination functions. Each way has two variants.
8191 Much of the structure is common to all four variations.
8193 @findex __CTOR_LIST__
8194 @findex __DTOR_LIST__
8195 The linker must build two lists of these functions---a list of
8196 initialization functions, called @code{__CTOR_LIST__}, and a list of
8197 termination functions, called @code{__DTOR_LIST__}.
8199 Each list always begins with an ignored function pointer (which may hold
8200 0, @minus{}1, or a count of the function pointers after it, depending on
8201 the environment). This is followed by a series of zero or more function
8202 pointers to constructors (or destructors), followed by a function
8203 pointer containing zero.
8205 Depending on the operating system and its executable file format, either
8206 @file{crtstuff.c} or @file{libgcc2.c} traverses these lists at startup
8207 time and exit time. Constructors are called in reverse order of the
8208 list; destructors in forward order.
8210 The best way to handle static constructors works only for object file
8211 formats which provide arbitrarily-named sections. A section is set
8212 aside for a list of constructors, and another for a list of destructors.
8213 Traditionally these are called @samp{.ctors} and @samp{.dtors}. Each
8214 object file that defines an initialization function also puts a word in
8215 the constructor section to point to that function. The linker
8216 accumulates all these words into one contiguous @samp{.ctors} section.
8217 Termination functions are handled similarly.
8219 This method will be chosen as the default by @file{target-def.h} if
8220 @code{TARGET_ASM_NAMED_SECTION} is defined. A target that does not
8221 support arbitrary sections, but does support special designated
8222 constructor and destructor sections may define @code{CTORS_SECTION_ASM_OP}
8223 and @code{DTORS_SECTION_ASM_OP} to achieve the same effect.
8225 When arbitrary sections are available, there are two variants, depending
8226 upon how the code in @file{crtstuff.c} is called. On systems that
8227 support a @dfn{.init} section which is executed at program startup,
8228 parts of @file{crtstuff.c} are compiled into that section. The
8229 program is linked by the @command{gcc} driver like this:
8232 ld -o @var{output_file} crti.o crtbegin.o @dots{} -lgcc crtend.o crtn.o
8235 The prologue of a function (@code{__init}) appears in the @code{.init}
8236 section of @file{crti.o}; the epilogue appears in @file{crtn.o}. Likewise
8237 for the function @code{__fini} in the @dfn{.fini} section. Normally these
8238 files are provided by the operating system or by the GNU C library, but
8239 are provided by GCC for a few targets.
8241 The objects @file{crtbegin.o} and @file{crtend.o} are (for most targets)
8242 compiled from @file{crtstuff.c}. They contain, among other things, code
8243 fragments within the @code{.init} and @code{.fini} sections that branch
8244 to routines in the @code{.text} section. The linker will pull all parts
8245 of a section together, which results in a complete @code{__init} function
8246 that invokes the routines we need at startup.
8248 To use this variant, you must define the @code{INIT_SECTION_ASM_OP}
8251 If no init section is available, when GCC compiles any function called
8252 @code{main} (or more accurately, any function designated as a program
8253 entry point by the language front end calling @code{expand_main_function}),
8254 it inserts a procedure call to @code{__main} as the first executable code
8255 after the function prologue. The @code{__main} function is defined
8256 in @file{libgcc2.c} and runs the global constructors.
8258 In file formats that don't support arbitrary sections, there are again
8259 two variants. In the simplest variant, the GNU linker (GNU @code{ld})
8260 and an `a.out' format must be used. In this case,
8261 @code{TARGET_ASM_CONSTRUCTOR} is defined to produce a @code{.stabs}
8262 entry of type @samp{N_SETT}, referencing the name @code{__CTOR_LIST__},
8263 and with the address of the void function containing the initialization
8264 code as its value. The GNU linker recognizes this as a request to add
8265 the value to a @dfn{set}; the values are accumulated, and are eventually
8266 placed in the executable as a vector in the format described above, with
8267 a leading (ignored) count and a trailing zero element.
8268 @code{TARGET_ASM_DESTRUCTOR} is handled similarly. Since no init
8269 section is available, the absence of @code{INIT_SECTION_ASM_OP} causes
8270 the compilation of @code{main} to call @code{__main} as above, starting
8271 the initialization process.
8273 The last variant uses neither arbitrary sections nor the GNU linker.
8274 This is preferable when you want to do dynamic linking and when using
8275 file formats which the GNU linker does not support, such as `ECOFF'@. In
8276 this case, @code{TARGET_HAVE_CTORS_DTORS} is false, initialization and
8277 termination functions are recognized simply by their names. This requires
8278 an extra program in the linkage step, called @command{collect2}. This program
8279 pretends to be the linker, for use with GCC; it does its job by running
8280 the ordinary linker, but also arranges to include the vectors of
8281 initialization and termination functions. These functions are called
8282 via @code{__main} as described above. In order to use this method,
8283 @code{use_collect2} must be defined in the target in @file{config.gcc}.
8286 The following section describes the specific macros that control and
8287 customize the handling of initialization and termination functions.
8290 @node Macros for Initialization
8291 @subsection Macros Controlling Initialization Routines
8293 Here are the macros that control how the compiler handles initialization
8294 and termination functions:
8296 @defmac INIT_SECTION_ASM_OP
8297 If defined, a C string constant, including spacing, for the assembler
8298 operation to identify the following data as initialization code. If not
8299 defined, GCC will assume such a section does not exist. When you are
8300 using special sections for initialization and termination functions, this
8301 macro also controls how @file{crtstuff.c} and @file{libgcc2.c} arrange to
8302 run the initialization functions.
8305 @defmac HAS_INIT_SECTION
8306 If defined, @code{main} will not call @code{__main} as described above.
8307 This macro should be defined for systems that control start-up code
8308 on a symbol-by-symbol basis, such as OSF/1, and should not
8309 be defined explicitly for systems that support @code{INIT_SECTION_ASM_OP}.
8312 @defmac LD_INIT_SWITCH
8313 If defined, a C string constant for a switch that tells the linker that
8314 the following symbol is an initialization routine.
8317 @defmac LD_FINI_SWITCH
8318 If defined, a C string constant for a switch that tells the linker that
8319 the following symbol is a finalization routine.
8322 @defmac COLLECT_SHARED_INIT_FUNC (@var{stream}, @var{func})
8323 If defined, a C statement that will write a function that can be
8324 automatically called when a shared library is loaded. The function
8325 should call @var{func}, which takes no arguments. If not defined, and
8326 the object format requires an explicit initialization function, then a
8327 function called @code{_GLOBAL__DI} will be generated.
8329 This function and the following one are used by collect2 when linking a
8330 shared library that needs constructors or destructors, or has DWARF2
8331 exception tables embedded in the code.
8334 @defmac COLLECT_SHARED_FINI_FUNC (@var{stream}, @var{func})
8335 If defined, a C statement that will write a function that can be
8336 automatically called when a shared library is unloaded. The function
8337 should call @var{func}, which takes no arguments. If not defined, and
8338 the object format requires an explicit finalization function, then a
8339 function called @code{_GLOBAL__DD} will be generated.
8342 @defmac INVOKE__main
8343 If defined, @code{main} will call @code{__main} despite the presence of
8344 @code{INIT_SECTION_ASM_OP}. This macro should be defined for systems
8345 where the init section is not actually run automatically, but is still
8346 useful for collecting the lists of constructors and destructors.
8349 @defmac SUPPORTS_INIT_PRIORITY
8350 If nonzero, the C++ @code{init_priority} attribute is supported and the
8351 compiler should emit instructions to control the order of initialization
8352 of objects. If zero, the compiler will issue an error message upon
8353 encountering an @code{init_priority} attribute.
8356 @deftypevr {Target Hook} bool TARGET_HAVE_CTORS_DTORS
8357 This value is true if the target supports some ``native'' method of
8358 collecting constructors and destructors to be run at startup and exit.
8359 It is false if we must use @command{collect2}.
8362 @deftypefn {Target Hook} void TARGET_ASM_CONSTRUCTOR (rtx @var{symbol}, int @var{priority})
8363 If defined, a function that outputs assembler code to arrange to call
8364 the function referenced by @var{symbol} at initialization time.
8366 Assume that @var{symbol} is a @code{SYMBOL_REF} for a function taking
8367 no arguments and with no return value. If the target supports initialization
8368 priorities, @var{priority} is a value between 0 and @code{MAX_INIT_PRIORITY};
8369 otherwise it must be @code{DEFAULT_INIT_PRIORITY}.
8371 If this macro is not defined by the target, a suitable default will
8372 be chosen if (1) the target supports arbitrary section names, (2) the
8373 target defines @code{CTORS_SECTION_ASM_OP}, or (3) @code{USE_COLLECT2}
8377 @deftypefn {Target Hook} void TARGET_ASM_DESTRUCTOR (rtx @var{symbol}, int @var{priority})
8378 This is like @code{TARGET_ASM_CONSTRUCTOR} but used for termination
8379 functions rather than initialization functions.
8382 If @code{TARGET_HAVE_CTORS_DTORS} is true, the initialization routine
8383 generated for the generated object file will have static linkage.
8385 If your system uses @command{collect2} as the means of processing
8386 constructors, then that program normally uses @command{nm} to scan
8387 an object file for constructor functions to be called.
8389 On certain kinds of systems, you can define this macro to make
8390 @command{collect2} work faster (and, in some cases, make it work at all):
8392 @defmac OBJECT_FORMAT_COFF
8393 Define this macro if the system uses COFF (Common Object File Format)
8394 object files, so that @command{collect2} can assume this format and scan
8395 object files directly for dynamic constructor/destructor functions.
8397 This macro is effective only in a native compiler; @command{collect2} as
8398 part of a cross compiler always uses @command{nm} for the target machine.
8401 @defmac REAL_NM_FILE_NAME
8402 Define this macro as a C string constant containing the file name to use
8403 to execute @command{nm}. The default is to search the path normally for
8406 If your system supports shared libraries and has a program to list the
8407 dynamic dependencies of a given library or executable, you can define
8408 these macros to enable support for running initialization and
8409 termination functions in shared libraries:
8413 Define this macro to a C string constant containing the name of the program
8414 which lists dynamic dependencies, like @command{"ldd"} under SunOS 4.
8417 @defmac PARSE_LDD_OUTPUT (@var{ptr})
8418 Define this macro to be C code that extracts filenames from the output
8419 of the program denoted by @code{LDD_SUFFIX}. @var{ptr} is a variable
8420 of type @code{char *} that points to the beginning of a line of output
8421 from @code{LDD_SUFFIX}. If the line lists a dynamic dependency, the
8422 code must advance @var{ptr} to the beginning of the filename on that
8423 line. Otherwise, it must set @var{ptr} to @code{NULL}.
8426 @defmac SHLIB_SUFFIX
8427 Define this macro to a C string constant containing the default shared
8428 library extension of the target (e.g., @samp{".so"}). @command{collect2}
8429 strips version information after this suffix when generating global
8430 constructor and destructor names. This define is only needed on targets
8431 that use @command{collect2} to process constructors and destructors.
8434 @node Instruction Output
8435 @subsection Output of Assembler Instructions
8437 @c prevent bad page break with this line
8438 This describes assembler instruction output.
8440 @defmac REGISTER_NAMES
8441 A C initializer containing the assembler's names for the machine
8442 registers, each one as a C string constant. This is what translates
8443 register numbers in the compiler into assembler language.
8446 @defmac ADDITIONAL_REGISTER_NAMES
8447 If defined, a C initializer for an array of structures containing a name
8448 and a register number. This macro defines additional names for hard
8449 registers, thus allowing the @code{asm} option in declarations to refer
8450 to registers using alternate names.
8453 @defmac ASM_OUTPUT_OPCODE (@var{stream}, @var{ptr})
8454 Define this macro if you are using an unusual assembler that
8455 requires different names for the machine instructions.
8457 The definition is a C statement or statements which output an
8458 assembler instruction opcode to the stdio stream @var{stream}. The
8459 macro-operand @var{ptr} is a variable of type @code{char *} which
8460 points to the opcode name in its ``internal'' form---the form that is
8461 written in the machine description. The definition should output the
8462 opcode name to @var{stream}, performing any translation you desire, and
8463 increment the variable @var{ptr} to point at the end of the opcode
8464 so that it will not be output twice.
8466 In fact, your macro definition may process less than the entire opcode
8467 name, or more than the opcode name; but if you want to process text
8468 that includes @samp{%}-sequences to substitute operands, you must take
8469 care of the substitution yourself. Just be sure to increment
8470 @var{ptr} over whatever text should not be output normally.
8472 @findex recog_data.operand
8473 If you need to look at the operand values, they can be found as the
8474 elements of @code{recog_data.operand}.
8476 If the macro definition does nothing, the instruction is output
8480 @defmac FINAL_PRESCAN_INSN (@var{insn}, @var{opvec}, @var{noperands})
8481 If defined, a C statement to be executed just prior to the output of
8482 assembler code for @var{insn}, to modify the extracted operands so
8483 they will be output differently.
8485 Here the argument @var{opvec} is the vector containing the operands
8486 extracted from @var{insn}, and @var{noperands} is the number of
8487 elements of the vector which contain meaningful data for this insn.
8488 The contents of this vector are what will be used to convert the insn
8489 template into assembler code, so you can change the assembler output
8490 by changing the contents of the vector.
8492 This macro is useful when various assembler syntaxes share a single
8493 file of instruction patterns; by defining this macro differently, you
8494 can cause a large class of instructions to be output differently (such
8495 as with rearranged operands). Naturally, variations in assembler
8496 syntax affecting individual insn patterns ought to be handled by
8497 writing conditional output routines in those patterns.
8499 If this macro is not defined, it is equivalent to a null statement.
8502 @deftypefn {Target Hook} void TARGET_ASM_FINAL_POSTSCAN_INSN (FILE *@var{file}, rtx @var{insn}, rtx *@var{opvec}, int @var{noperands})
8503 If defined, this target hook is a function which is executed just after the
8504 output of assembler code for @var{insn}, to change the mode of the assembler
8507 Here the argument @var{opvec} is the vector containing the operands
8508 extracted from @var{insn}, and @var{noperands} is the number of
8509 elements of the vector which contain meaningful data for this insn.
8510 The contents of this vector are what was used to convert the insn
8511 template into assembler code, so you can change the assembler mode
8512 by checking the contents of the vector.
8515 @defmac PRINT_OPERAND (@var{stream}, @var{x}, @var{code})
8516 A C compound statement to output to stdio stream @var{stream} the
8517 assembler syntax for an instruction operand @var{x}. @var{x} is an
8520 @var{code} is a value that can be used to specify one of several ways
8521 of printing the operand. It is used when identical operands must be
8522 printed differently depending on the context. @var{code} comes from
8523 the @samp{%} specification that was used to request printing of the
8524 operand. If the specification was just @samp{%@var{digit}} then
8525 @var{code} is 0; if the specification was @samp{%@var{ltr}
8526 @var{digit}} then @var{code} is the ASCII code for @var{ltr}.
8529 If @var{x} is a register, this macro should print the register's name.
8530 The names can be found in an array @code{reg_names} whose type is
8531 @code{char *[]}. @code{reg_names} is initialized from
8532 @code{REGISTER_NAMES}.
8534 When the machine description has a specification @samp{%@var{punct}}
8535 (a @samp{%} followed by a punctuation character), this macro is called
8536 with a null pointer for @var{x} and the punctuation character for
8540 @defmac PRINT_OPERAND_PUNCT_VALID_P (@var{code})
8541 A C expression which evaluates to true if @var{code} is a valid
8542 punctuation character for use in the @code{PRINT_OPERAND} macro. If
8543 @code{PRINT_OPERAND_PUNCT_VALID_P} is not defined, it means that no
8544 punctuation characters (except for the standard one, @samp{%}) are used
8548 @defmac PRINT_OPERAND_ADDRESS (@var{stream}, @var{x})
8549 A C compound statement to output to stdio stream @var{stream} the
8550 assembler syntax for an instruction operand that is a memory reference
8551 whose address is @var{x}. @var{x} is an RTL expression.
8553 @cindex @code{TARGET_ENCODE_SECTION_INFO} usage
8554 On some machines, the syntax for a symbolic address depends on the
8555 section that the address refers to. On these machines, define the hook
8556 @code{TARGET_ENCODE_SECTION_INFO} to store the information into the
8557 @code{symbol_ref}, and then check for it here. @xref{Assembler
8561 @findex dbr_sequence_length
8562 @defmac DBR_OUTPUT_SEQEND (@var{file})
8563 A C statement, to be executed after all slot-filler instructions have
8564 been output. If necessary, call @code{dbr_sequence_length} to
8565 determine the number of slots filled in a sequence (zero if not
8566 currently outputting a sequence), to decide how many no-ops to output,
8569 Don't define this macro if it has nothing to do, but it is helpful in
8570 reading assembly output if the extent of the delay sequence is made
8571 explicit (e.g.@: with white space).
8574 @findex final_sequence
8575 Note that output routines for instructions with delay slots must be
8576 prepared to deal with not being output as part of a sequence
8577 (i.e.@: when the scheduling pass is not run, or when no slot fillers could be
8578 found.) The variable @code{final_sequence} is null when not
8579 processing a sequence, otherwise it contains the @code{sequence} rtx
8583 @defmac REGISTER_PREFIX
8584 @defmacx LOCAL_LABEL_PREFIX
8585 @defmacx USER_LABEL_PREFIX
8586 @defmacx IMMEDIATE_PREFIX
8587 If defined, C string expressions to be used for the @samp{%R}, @samp{%L},
8588 @samp{%U}, and @samp{%I} options of @code{asm_fprintf} (see
8589 @file{final.c}). These are useful when a single @file{md} file must
8590 support multiple assembler formats. In that case, the various @file{tm.h}
8591 files can define these macros differently.
8594 @defmac ASM_FPRINTF_EXTENSIONS (@var{file}, @var{argptr}, @var{format})
8595 If defined this macro should expand to a series of @code{case}
8596 statements which will be parsed inside the @code{switch} statement of
8597 the @code{asm_fprintf} function. This allows targets to define extra
8598 printf formats which may useful when generating their assembler
8599 statements. Note that uppercase letters are reserved for future
8600 generic extensions to asm_fprintf, and so are not available to target
8601 specific code. The output file is given by the parameter @var{file}.
8602 The varargs input pointer is @var{argptr} and the rest of the format
8603 string, starting the character after the one that is being switched
8604 upon, is pointed to by @var{format}.
8607 @defmac ASSEMBLER_DIALECT
8608 If your target supports multiple dialects of assembler language (such as
8609 different opcodes), define this macro as a C expression that gives the
8610 numeric index of the assembler language dialect to use, with zero as the
8613 If this macro is defined, you may use constructs of the form
8615 @samp{@{option0|option1|option2@dots{}@}}
8618 in the output templates of patterns (@pxref{Output Template}) or in the
8619 first argument of @code{asm_fprintf}. This construct outputs
8620 @samp{option0}, @samp{option1}, @samp{option2}, etc., if the value of
8621 @code{ASSEMBLER_DIALECT} is zero, one, two, etc. Any special characters
8622 within these strings retain their usual meaning. If there are fewer
8623 alternatives within the braces than the value of
8624 @code{ASSEMBLER_DIALECT}, the construct outputs nothing.
8626 If you do not define this macro, the characters @samp{@{}, @samp{|} and
8627 @samp{@}} do not have any special meaning when used in templates or
8628 operands to @code{asm_fprintf}.
8630 Define the macros @code{REGISTER_PREFIX}, @code{LOCAL_LABEL_PREFIX},
8631 @code{USER_LABEL_PREFIX} and @code{IMMEDIATE_PREFIX} if you can express
8632 the variations in assembler language syntax with that mechanism. Define
8633 @code{ASSEMBLER_DIALECT} and use the @samp{@{option0|option1@}} syntax
8634 if the syntax variant are larger and involve such things as different
8635 opcodes or operand order.
8638 @defmac ASM_OUTPUT_REG_PUSH (@var{stream}, @var{regno})
8639 A C expression to output to @var{stream} some assembler code
8640 which will push hard register number @var{regno} onto the stack.
8641 The code need not be optimal, since this macro is used only when
8645 @defmac ASM_OUTPUT_REG_POP (@var{stream}, @var{regno})
8646 A C expression to output to @var{stream} some assembler code
8647 which will pop hard register number @var{regno} off of the stack.
8648 The code need not be optimal, since this macro is used only when
8652 @node Dispatch Tables
8653 @subsection Output of Dispatch Tables
8655 @c prevent bad page break with this line
8656 This concerns dispatch tables.
8658 @cindex dispatch table
8659 @defmac ASM_OUTPUT_ADDR_DIFF_ELT (@var{stream}, @var{body}, @var{value}, @var{rel})
8660 A C statement to output to the stdio stream @var{stream} an assembler
8661 pseudo-instruction to generate a difference between two labels.
8662 @var{value} and @var{rel} are the numbers of two internal labels. The
8663 definitions of these labels are output using
8664 @code{(*targetm.asm_out.internal_label)}, and they must be printed in the same
8665 way here. For example,
8668 fprintf (@var{stream}, "\t.word L%d-L%d\n",
8669 @var{value}, @var{rel})
8672 You must provide this macro on machines where the addresses in a
8673 dispatch table are relative to the table's own address. If defined, GCC
8674 will also use this macro on all machines when producing PIC@.
8675 @var{body} is the body of the @code{ADDR_DIFF_VEC}; it is provided so that the
8676 mode and flags can be read.
8679 @defmac ASM_OUTPUT_ADDR_VEC_ELT (@var{stream}, @var{value})
8680 This macro should be provided on machines where the addresses
8681 in a dispatch table are absolute.
8683 The definition should be a C statement to output to the stdio stream
8684 @var{stream} an assembler pseudo-instruction to generate a reference to
8685 a label. @var{value} is the number of an internal label whose
8686 definition is output using @code{(*targetm.asm_out.internal_label)}.
8690 fprintf (@var{stream}, "\t.word L%d\n", @var{value})
8694 @defmac ASM_OUTPUT_CASE_LABEL (@var{stream}, @var{prefix}, @var{num}, @var{table})
8695 Define this if the label before a jump-table needs to be output
8696 specially. The first three arguments are the same as for
8697 @code{(*targetm.asm_out.internal_label)}; the fourth argument is the
8698 jump-table which follows (a @code{jump_insn} containing an
8699 @code{addr_vec} or @code{addr_diff_vec}).
8701 This feature is used on system V to output a @code{swbeg} statement
8704 If this macro is not defined, these labels are output with
8705 @code{(*targetm.asm_out.internal_label)}.
8708 @defmac ASM_OUTPUT_CASE_END (@var{stream}, @var{num}, @var{table})
8709 Define this if something special must be output at the end of a
8710 jump-table. The definition should be a C statement to be executed
8711 after the assembler code for the table is written. It should write
8712 the appropriate code to stdio stream @var{stream}. The argument
8713 @var{table} is the jump-table insn, and @var{num} is the label-number
8714 of the preceding label.
8716 If this macro is not defined, nothing special is output at the end of
8720 @deftypefn {Target Hook} void TARGET_ASM_EMIT_UNWIND_LABEL (FILE *@var{stream}, tree @var{decl}, int @var{for_eh}, int @var{empty})
8721 This target hook emits a label at the beginning of each FDE@. It
8722 should be defined on targets where FDEs need special labels, and it
8723 should write the appropriate label, for the FDE associated with the
8724 function declaration @var{decl}, to the stdio stream @var{stream}.
8725 The third argument, @var{for_eh}, is a boolean: true if this is for an
8726 exception table. The fourth argument, @var{empty}, is a boolean:
8727 true if this is a placeholder label for an omitted FDE@.
8729 The default is that FDEs are not given nonlocal labels.
8732 @deftypefn {Target Hook} void TARGET_ASM_EMIT_EXCEPT_TABLE_LABEL (FILE *@var{stream})
8733 This target hook emits a label at the beginning of the exception table.
8734 It should be defined on targets where it is desirable for the table
8735 to be broken up according to function.
8737 The default is that no label is emitted.
8740 @deftypefn {Target Hook} void TARGET_ASM_EMIT_EXCEPT_PERSONALITY (rtx @var{personality})
8741 If the target implements @code{TARGET_ASM_UNWIND_EMIT}, this hook may be used to emit a directive to install a personality hook into the unwind info. This hook should not be used if dwarf2 unwind info is used.
8744 @deftypefn {Target Hook} void TARGET_ASM_UNWIND_EMIT (FILE *@var{stream}, rtx @var{insn})
8745 This target hook emits assembly directives required to unwind the
8746 given instruction. This is only used when TARGET_UNWIND_INFO is set.
8749 @deftypevr {Target Hook} bool TARGET_ASM_UNWIND_EMIT_BEFORE_INSN
8750 True if the @code{TARGET_ASM_UNWIND_EMIT} hook should be called before the assembly for @var{insn} has been emitted, false if the hook should be called afterward.
8753 @node Exception Region Output
8754 @subsection Assembler Commands for Exception Regions
8756 @c prevent bad page break with this line
8758 This describes commands marking the start and the end of an exception
8761 @defmac EH_FRAME_SECTION_NAME
8762 If defined, a C string constant for the name of the section containing
8763 exception handling frame unwind information. If not defined, GCC will
8764 provide a default definition if the target supports named sections.
8765 @file{crtstuff.c} uses this macro to switch to the appropriate section.
8767 You should define this symbol if your target supports DWARF 2 frame
8768 unwind information and the default definition does not work.
8771 @defmac EH_FRAME_IN_DATA_SECTION
8772 If defined, DWARF 2 frame unwind information will be placed in the
8773 data section even though the target supports named sections. This
8774 might be necessary, for instance, if the system linker does garbage
8775 collection and sections cannot be marked as not to be collected.
8777 Do not define this macro unless @code{TARGET_ASM_NAMED_SECTION} is
8781 @defmac EH_TABLES_CAN_BE_READ_ONLY
8782 Define this macro to 1 if your target is such that no frame unwind
8783 information encoding used with non-PIC code will ever require a
8784 runtime relocation, but the linker may not support merging read-only
8785 and read-write sections into a single read-write section.
8788 @defmac MASK_RETURN_ADDR
8789 An rtx used to mask the return address found via @code{RETURN_ADDR_RTX}, so
8790 that it does not contain any extraneous set bits in it.
8793 @defmac DWARF2_UNWIND_INFO
8794 Define this macro to 0 if your target supports DWARF 2 frame unwind
8795 information, but it does not yet work with exception handling.
8796 Otherwise, if your target supports this information (if it defines
8797 @samp{INCOMING_RETURN_ADDR_RTX} and either @samp{UNALIGNED_INT_ASM_OP}
8798 or @samp{OBJECT_FORMAT_ELF}), GCC will provide a default definition of 1.
8800 If @code{TARGET_UNWIND_INFO} is defined, the target specific unwinder
8801 will be used in all cases. Defining this macro will enable the generation
8802 of DWARF 2 frame debugging information.
8804 If @code{TARGET_UNWIND_INFO} is not defined, and this macro is defined to 1,
8805 the DWARF 2 unwinder will be the default exception handling mechanism;
8806 otherwise, the @code{setjmp}/@code{longjmp}-based scheme will be used by
8810 @defmac TARGET_UNWIND_INFO
8811 Define this macro if your target has ABI specified unwind tables. Usually
8812 these will be output by @code{TARGET_ASM_UNWIND_EMIT}.
8815 @deftypevr {Target Hook} bool TARGET_UNWIND_TABLES_DEFAULT
8816 This variable should be set to @code{true} if the target ABI requires unwinding
8817 tables even when exceptions are not used.
8820 @defmac MUST_USE_SJLJ_EXCEPTIONS
8821 This macro need only be defined if @code{DWARF2_UNWIND_INFO} is
8822 runtime-variable. In that case, @file{except.h} cannot correctly
8823 determine the corresponding definition of @code{MUST_USE_SJLJ_EXCEPTIONS},
8824 so the target must provide it directly.
8827 @defmac DONT_USE_BUILTIN_SETJMP
8828 Define this macro to 1 if the @code{setjmp}/@code{longjmp}-based scheme
8829 should use the @code{setjmp}/@code{longjmp} functions from the C library
8830 instead of the @code{__builtin_setjmp}/@code{__builtin_longjmp} machinery.
8833 @defmac DWARF_CIE_DATA_ALIGNMENT
8834 This macro need only be defined if the target might save registers in the
8835 function prologue at an offset to the stack pointer that is not aligned to
8836 @code{UNITS_PER_WORD}. The definition should be the negative minimum
8837 alignment if @code{STACK_GROWS_DOWNWARD} is defined, and the positive
8838 minimum alignment otherwise. @xref{SDB and DWARF}. Only applicable if
8839 the target supports DWARF 2 frame unwind information.
8842 @deftypevr {Target Hook} bool TARGET_TERMINATE_DW2_EH_FRAME_INFO
8843 Contains the value true if the target should add a zero word onto the
8844 end of a Dwarf-2 frame info section when used for exception handling.
8845 Default value is false if @code{EH_FRAME_SECTION_NAME} is defined, and
8849 @deftypefn {Target Hook} rtx TARGET_DWARF_REGISTER_SPAN (rtx @var{reg})
8850 Given a register, this hook should return a parallel of registers to
8851 represent where to find the register pieces. Define this hook if the
8852 register and its mode are represented in Dwarf in non-contiguous
8853 locations, or if the register should be represented in more than one
8854 register in Dwarf. Otherwise, this hook should return @code{NULL_RTX}.
8855 If not defined, the default is to return @code{NULL_RTX}.
8858 @deftypefn {Target Hook} void TARGET_INIT_DWARF_REG_SIZES_EXTRA (tree @var{address})
8859 If some registers are represented in Dwarf-2 unwind information in
8860 multiple pieces, define this hook to fill in information about the
8861 sizes of those pieces in the table used by the unwinder at runtime.
8862 It will be called by @code{expand_builtin_init_dwarf_reg_sizes} after
8863 filling in a single size corresponding to each hard register;
8864 @var{address} is the address of the table.
8867 @deftypefn {Target Hook} bool TARGET_ASM_TTYPE (rtx @var{sym})
8868 This hook is used to output a reference from a frame unwinding table to
8869 the type_info object identified by @var{sym}. It should return @code{true}
8870 if the reference was output. Returning @code{false} will cause the
8871 reference to be output using the normal Dwarf2 routines.
8874 @deftypevr {Target Hook} bool TARGET_ARM_EABI_UNWINDER
8875 This flag should be set to @code{true} on targets that use an ARM EABI
8876 based unwinding library, and @code{false} on other targets. This effects
8877 the format of unwinding tables, and how the unwinder in entered after
8878 running a cleanup. The default is @code{false}.
8881 @node Alignment Output
8882 @subsection Assembler Commands for Alignment
8884 @c prevent bad page break with this line
8885 This describes commands for alignment.
8887 @defmac JUMP_ALIGN (@var{label})
8888 The alignment (log base 2) to put in front of @var{label}, which is
8889 a common destination of jumps and has no fallthru incoming edge.
8891 This macro need not be defined if you don't want any special alignment
8892 to be done at such a time. Most machine descriptions do not currently
8895 Unless it's necessary to inspect the @var{label} parameter, it is better
8896 to set the variable @var{align_jumps} in the target's
8897 @code{TARGET_OPTION_OVERRIDE}. Otherwise, you should try to honor the user's
8898 selection in @var{align_jumps} in a @code{JUMP_ALIGN} implementation.
8901 @defmac LABEL_ALIGN_AFTER_BARRIER (@var{label})
8902 The alignment (log base 2) to put in front of @var{label}, which follows
8905 This macro need not be defined if you don't want any special alignment
8906 to be done at such a time. Most machine descriptions do not currently
8910 @defmac LABEL_ALIGN_AFTER_BARRIER_MAX_SKIP
8911 The maximum number of bytes to skip when applying
8912 @code{LABEL_ALIGN_AFTER_BARRIER}. This works only if
8913 @code{ASM_OUTPUT_MAX_SKIP_ALIGN} is defined.
8916 @defmac LOOP_ALIGN (@var{label})
8917 The alignment (log base 2) to put in front of @var{label}, which follows
8918 a @code{NOTE_INSN_LOOP_BEG} note.
8920 This macro need not be defined if you don't want any special alignment
8921 to be done at such a time. Most machine descriptions do not currently
8924 Unless it's necessary to inspect the @var{label} parameter, it is better
8925 to set the variable @code{align_loops} in the target's
8926 @code{TARGET_OPTION_OVERRIDE}. Otherwise, you should try to honor the user's
8927 selection in @code{align_loops} in a @code{LOOP_ALIGN} implementation.
8930 @defmac LOOP_ALIGN_MAX_SKIP
8931 The maximum number of bytes to skip when applying @code{LOOP_ALIGN}.
8932 This works only if @code{ASM_OUTPUT_MAX_SKIP_ALIGN} is defined.
8935 @defmac LABEL_ALIGN (@var{label})
8936 The alignment (log base 2) to put in front of @var{label}.
8937 If @code{LABEL_ALIGN_AFTER_BARRIER} / @code{LOOP_ALIGN} specify a different alignment,
8938 the maximum of the specified values is used.
8940 Unless it's necessary to inspect the @var{label} parameter, it is better
8941 to set the variable @code{align_labels} in the target's
8942 @code{TARGET_OPTION_OVERRIDE}. Otherwise, you should try to honor the user's
8943 selection in @code{align_labels} in a @code{LABEL_ALIGN} implementation.
8946 @defmac LABEL_ALIGN_MAX_SKIP
8947 The maximum number of bytes to skip when applying @code{LABEL_ALIGN}.
8948 This works only if @code{ASM_OUTPUT_MAX_SKIP_ALIGN} is defined.
8951 @defmac ASM_OUTPUT_SKIP (@var{stream}, @var{nbytes})
8952 A C statement to output to the stdio stream @var{stream} an assembler
8953 instruction to advance the location counter by @var{nbytes} bytes.
8954 Those bytes should be zero when loaded. @var{nbytes} will be a C
8955 expression of type @code{unsigned HOST_WIDE_INT}.
8958 @defmac ASM_NO_SKIP_IN_TEXT
8959 Define this macro if @code{ASM_OUTPUT_SKIP} should not be used in the
8960 text section because it fails to put zeros in the bytes that are skipped.
8961 This is true on many Unix systems, where the pseudo--op to skip bytes
8962 produces no-op instructions rather than zeros when used in the text
8966 @defmac ASM_OUTPUT_ALIGN (@var{stream}, @var{power})
8967 A C statement to output to the stdio stream @var{stream} an assembler
8968 command to advance the location counter to a multiple of 2 to the
8969 @var{power} bytes. @var{power} will be a C expression of type @code{int}.
8972 @defmac ASM_OUTPUT_ALIGN_WITH_NOP (@var{stream}, @var{power})
8973 Like @code{ASM_OUTPUT_ALIGN}, except that the ``nop'' instruction is used
8974 for padding, if necessary.
8977 @defmac ASM_OUTPUT_MAX_SKIP_ALIGN (@var{stream}, @var{power}, @var{max_skip})
8978 A C statement to output to the stdio stream @var{stream} an assembler
8979 command to advance the location counter to a multiple of 2 to the
8980 @var{power} bytes, but only if @var{max_skip} or fewer bytes are needed to
8981 satisfy the alignment request. @var{power} and @var{max_skip} will be
8982 a C expression of type @code{int}.
8986 @node Debugging Info
8987 @section Controlling Debugging Information Format
8989 @c prevent bad page break with this line
8990 This describes how to specify debugging information.
8993 * All Debuggers:: Macros that affect all debugging formats uniformly.
8994 * DBX Options:: Macros enabling specific options in DBX format.
8995 * DBX Hooks:: Hook macros for varying DBX format.
8996 * File Names and DBX:: Macros controlling output of file names in DBX format.
8997 * SDB and DWARF:: Macros for SDB (COFF) and DWARF formats.
8998 * VMS Debug:: Macros for VMS debug format.
9002 @subsection Macros Affecting All Debugging Formats
9004 @c prevent bad page break with this line
9005 These macros affect all debugging formats.
9007 @defmac DBX_REGISTER_NUMBER (@var{regno})
9008 A C expression that returns the DBX register number for the compiler
9009 register number @var{regno}. In the default macro provided, the value
9010 of this expression will be @var{regno} itself. But sometimes there are
9011 some registers that the compiler knows about and DBX does not, or vice
9012 versa. In such cases, some register may need to have one number in the
9013 compiler and another for DBX@.
9015 If two registers have consecutive numbers inside GCC, and they can be
9016 used as a pair to hold a multiword value, then they @emph{must} have
9017 consecutive numbers after renumbering with @code{DBX_REGISTER_NUMBER}.
9018 Otherwise, debuggers will be unable to access such a pair, because they
9019 expect register pairs to be consecutive in their own numbering scheme.
9021 If you find yourself defining @code{DBX_REGISTER_NUMBER} in way that
9022 does not preserve register pairs, then what you must do instead is
9023 redefine the actual register numbering scheme.
9026 @defmac DEBUGGER_AUTO_OFFSET (@var{x})
9027 A C expression that returns the integer offset value for an automatic
9028 variable having address @var{x} (an RTL expression). The default
9029 computation assumes that @var{x} is based on the frame-pointer and
9030 gives the offset from the frame-pointer. This is required for targets
9031 that produce debugging output for DBX or COFF-style debugging output
9032 for SDB and allow the frame-pointer to be eliminated when the
9033 @option{-g} options is used.
9036 @defmac DEBUGGER_ARG_OFFSET (@var{offset}, @var{x})
9037 A C expression that returns the integer offset value for an argument
9038 having address @var{x} (an RTL expression). The nominal offset is
9042 @defmac PREFERRED_DEBUGGING_TYPE
9043 A C expression that returns the type of debugging output GCC should
9044 produce when the user specifies just @option{-g}. Define
9045 this if you have arranged for GCC to support more than one format of
9046 debugging output. Currently, the allowable values are @code{DBX_DEBUG},
9047 @code{SDB_DEBUG}, @code{DWARF_DEBUG}, @code{DWARF2_DEBUG},
9048 @code{XCOFF_DEBUG}, @code{VMS_DEBUG}, and @code{VMS_AND_DWARF2_DEBUG}.
9050 When the user specifies @option{-ggdb}, GCC normally also uses the
9051 value of this macro to select the debugging output format, but with two
9052 exceptions. If @code{DWARF2_DEBUGGING_INFO} is defined, GCC uses the
9053 value @code{DWARF2_DEBUG}. Otherwise, if @code{DBX_DEBUGGING_INFO} is
9054 defined, GCC uses @code{DBX_DEBUG}.
9056 The value of this macro only affects the default debugging output; the
9057 user can always get a specific type of output by using @option{-gstabs},
9058 @option{-gcoff}, @option{-gdwarf-2}, @option{-gxcoff}, or @option{-gvms}.
9062 @subsection Specific Options for DBX Output
9064 @c prevent bad page break with this line
9065 These are specific options for DBX output.
9067 @defmac DBX_DEBUGGING_INFO
9068 Define this macro if GCC should produce debugging output for DBX
9069 in response to the @option{-g} option.
9072 @defmac XCOFF_DEBUGGING_INFO
9073 Define this macro if GCC should produce XCOFF format debugging output
9074 in response to the @option{-g} option. This is a variant of DBX format.
9077 @defmac DEFAULT_GDB_EXTENSIONS
9078 Define this macro to control whether GCC should by default generate
9079 GDB's extended version of DBX debugging information (assuming DBX-format
9080 debugging information is enabled at all). If you don't define the
9081 macro, the default is 1: always generate the extended information
9082 if there is any occasion to.
9085 @defmac DEBUG_SYMS_TEXT
9086 Define this macro if all @code{.stabs} commands should be output while
9087 in the text section.
9090 @defmac ASM_STABS_OP
9091 A C string constant, including spacing, naming the assembler pseudo op to
9092 use instead of @code{"\t.stabs\t"} to define an ordinary debugging symbol.
9093 If you don't define this macro, @code{"\t.stabs\t"} is used. This macro
9094 applies only to DBX debugging information format.
9097 @defmac ASM_STABD_OP
9098 A C string constant, including spacing, naming the assembler pseudo op to
9099 use instead of @code{"\t.stabd\t"} to define a debugging symbol whose
9100 value is the current location. If you don't define this macro,
9101 @code{"\t.stabd\t"} is used. This macro applies only to DBX debugging
9105 @defmac ASM_STABN_OP
9106 A C string constant, including spacing, naming the assembler pseudo op to
9107 use instead of @code{"\t.stabn\t"} to define a debugging symbol with no
9108 name. If you don't define this macro, @code{"\t.stabn\t"} is used. This
9109 macro applies only to DBX debugging information format.
9112 @defmac DBX_NO_XREFS
9113 Define this macro if DBX on your system does not support the construct
9114 @samp{xs@var{tagname}}. On some systems, this construct is used to
9115 describe a forward reference to a structure named @var{tagname}.
9116 On other systems, this construct is not supported at all.
9119 @defmac DBX_CONTIN_LENGTH
9120 A symbol name in DBX-format debugging information is normally
9121 continued (split into two separate @code{.stabs} directives) when it
9122 exceeds a certain length (by default, 80 characters). On some
9123 operating systems, DBX requires this splitting; on others, splitting
9124 must not be done. You can inhibit splitting by defining this macro
9125 with the value zero. You can override the default splitting-length by
9126 defining this macro as an expression for the length you desire.
9129 @defmac DBX_CONTIN_CHAR
9130 Normally continuation is indicated by adding a @samp{\} character to
9131 the end of a @code{.stabs} string when a continuation follows. To use
9132 a different character instead, define this macro as a character
9133 constant for the character you want to use. Do not define this macro
9134 if backslash is correct for your system.
9137 @defmac DBX_STATIC_STAB_DATA_SECTION
9138 Define this macro if it is necessary to go to the data section before
9139 outputting the @samp{.stabs} pseudo-op for a non-global static
9143 @defmac DBX_TYPE_DECL_STABS_CODE
9144 The value to use in the ``code'' field of the @code{.stabs} directive
9145 for a typedef. The default is @code{N_LSYM}.
9148 @defmac DBX_STATIC_CONST_VAR_CODE
9149 The value to use in the ``code'' field of the @code{.stabs} directive
9150 for a static variable located in the text section. DBX format does not
9151 provide any ``right'' way to do this. The default is @code{N_FUN}.
9154 @defmac DBX_REGPARM_STABS_CODE
9155 The value to use in the ``code'' field of the @code{.stabs} directive
9156 for a parameter passed in registers. DBX format does not provide any
9157 ``right'' way to do this. The default is @code{N_RSYM}.
9160 @defmac DBX_REGPARM_STABS_LETTER
9161 The letter to use in DBX symbol data to identify a symbol as a parameter
9162 passed in registers. DBX format does not customarily provide any way to
9163 do this. The default is @code{'P'}.
9166 @defmac DBX_FUNCTION_FIRST
9167 Define this macro if the DBX information for a function and its
9168 arguments should precede the assembler code for the function. Normally,
9169 in DBX format, the debugging information entirely follows the assembler
9173 @defmac DBX_BLOCKS_FUNCTION_RELATIVE
9174 Define this macro, with value 1, if the value of a symbol describing
9175 the scope of a block (@code{N_LBRAC} or @code{N_RBRAC}) should be
9176 relative to the start of the enclosing function. Normally, GCC uses
9177 an absolute address.
9180 @defmac DBX_LINES_FUNCTION_RELATIVE
9181 Define this macro, with value 1, if the value of a symbol indicating
9182 the current line number (@code{N_SLINE}) should be relative to the
9183 start of the enclosing function. Normally, GCC uses an absolute address.
9186 @defmac DBX_USE_BINCL
9187 Define this macro if GCC should generate @code{N_BINCL} and
9188 @code{N_EINCL} stabs for included header files, as on Sun systems. This
9189 macro also directs GCC to output a type number as a pair of a file
9190 number and a type number within the file. Normally, GCC does not
9191 generate @code{N_BINCL} or @code{N_EINCL} stabs, and it outputs a single
9192 number for a type number.
9196 @subsection Open-Ended Hooks for DBX Format
9198 @c prevent bad page break with this line
9199 These are hooks for DBX format.
9201 @defmac DBX_OUTPUT_LBRAC (@var{stream}, @var{name})
9202 Define this macro to say how to output to @var{stream} the debugging
9203 information for the start of a scope level for variable names. The
9204 argument @var{name} is the name of an assembler symbol (for use with
9205 @code{assemble_name}) whose value is the address where the scope begins.
9208 @defmac DBX_OUTPUT_RBRAC (@var{stream}, @var{name})
9209 Like @code{DBX_OUTPUT_LBRAC}, but for the end of a scope level.
9212 @defmac DBX_OUTPUT_NFUN (@var{stream}, @var{lscope_label}, @var{decl})
9213 Define this macro if the target machine requires special handling to
9214 output an @code{N_FUN} entry for the function @var{decl}.
9217 @defmac DBX_OUTPUT_SOURCE_LINE (@var{stream}, @var{line}, @var{counter})
9218 A C statement to output DBX debugging information before code for line
9219 number @var{line} of the current source file to the stdio stream
9220 @var{stream}. @var{counter} is the number of time the macro was
9221 invoked, including the current invocation; it is intended to generate
9222 unique labels in the assembly output.
9224 This macro should not be defined if the default output is correct, or
9225 if it can be made correct by defining @code{DBX_LINES_FUNCTION_RELATIVE}.
9228 @defmac NO_DBX_FUNCTION_END
9229 Some stabs encapsulation formats (in particular ECOFF), cannot handle the
9230 @code{.stabs "",N_FUN,,0,0,Lscope-function-1} gdb dbx extension construct.
9231 On those machines, define this macro to turn this feature off without
9232 disturbing the rest of the gdb extensions.
9235 @defmac NO_DBX_BNSYM_ENSYM
9236 Some assemblers cannot handle the @code{.stabd BNSYM/ENSYM,0,0} gdb dbx
9237 extension construct. On those machines, define this macro to turn this
9238 feature off without disturbing the rest of the gdb extensions.
9241 @node File Names and DBX
9242 @subsection File Names in DBX Format
9244 @c prevent bad page break with this line
9245 This describes file names in DBX format.
9247 @defmac DBX_OUTPUT_MAIN_SOURCE_FILENAME (@var{stream}, @var{name})
9248 A C statement to output DBX debugging information to the stdio stream
9249 @var{stream}, which indicates that file @var{name} is the main source
9250 file---the file specified as the input file for compilation.
9251 This macro is called only once, at the beginning of compilation.
9253 This macro need not be defined if the standard form of output
9254 for DBX debugging information is appropriate.
9256 It may be necessary to refer to a label equal to the beginning of the
9257 text section. You can use @samp{assemble_name (stream, ltext_label_name)}
9258 to do so. If you do this, you must also set the variable
9259 @var{used_ltext_label_name} to @code{true}.
9262 @defmac NO_DBX_MAIN_SOURCE_DIRECTORY
9263 Define this macro, with value 1, if GCC should not emit an indication
9264 of the current directory for compilation and current source language at
9265 the beginning of the file.
9268 @defmac NO_DBX_GCC_MARKER
9269 Define this macro, with value 1, if GCC should not emit an indication
9270 that this object file was compiled by GCC@. The default is to emit
9271 an @code{N_OPT} stab at the beginning of every source file, with
9272 @samp{gcc2_compiled.} for the string and value 0.
9275 @defmac DBX_OUTPUT_MAIN_SOURCE_FILE_END (@var{stream}, @var{name})
9276 A C statement to output DBX debugging information at the end of
9277 compilation of the main source file @var{name}. Output should be
9278 written to the stdio stream @var{stream}.
9280 If you don't define this macro, nothing special is output at the end
9281 of compilation, which is correct for most machines.
9284 @defmac DBX_OUTPUT_NULL_N_SO_AT_MAIN_SOURCE_FILE_END
9285 Define this macro @emph{instead of} defining
9286 @code{DBX_OUTPUT_MAIN_SOURCE_FILE_END}, if what needs to be output at
9287 the end of compilation is an @code{N_SO} stab with an empty string,
9288 whose value is the highest absolute text address in the file.
9293 @subsection Macros for SDB and DWARF Output
9295 @c prevent bad page break with this line
9296 Here are macros for SDB and DWARF output.
9298 @defmac SDB_DEBUGGING_INFO
9299 Define this macro if GCC should produce COFF-style debugging output
9300 for SDB in response to the @option{-g} option.
9303 @defmac DWARF2_DEBUGGING_INFO
9304 Define this macro if GCC should produce dwarf version 2 format
9305 debugging output in response to the @option{-g} option.
9307 @deftypefn {Target Hook} int TARGET_DWARF_CALLING_CONVENTION (const_tree @var{function})
9308 Define this to enable the dwarf attribute @code{DW_AT_calling_convention} to
9309 be emitted for each function. Instead of an integer return the enum
9310 value for the @code{DW_CC_} tag.
9313 To support optional call frame debugging information, you must also
9314 define @code{INCOMING_RETURN_ADDR_RTX} and either set
9315 @code{RTX_FRAME_RELATED_P} on the prologue insns if you use RTL for the
9316 prologue, or call @code{dwarf2out_def_cfa} and @code{dwarf2out_reg_save}
9317 as appropriate from @code{TARGET_ASM_FUNCTION_PROLOGUE} if you don't.
9320 @defmac DWARF2_FRAME_INFO
9321 Define this macro to a nonzero value if GCC should always output
9322 Dwarf 2 frame information. If @code{DWARF2_UNWIND_INFO}
9323 (@pxref{Exception Region Output} is nonzero, GCC will output this
9324 information not matter how you define @code{DWARF2_FRAME_INFO}.
9327 @defmac DWARF2_ASM_LINE_DEBUG_INFO
9328 Define this macro to be a nonzero value if the assembler can generate Dwarf 2
9329 line debug info sections. This will result in much more compact line number
9330 tables, and hence is desirable if it works.
9333 @deftypevr {Target Hook} bool TARGET_WANT_DEBUG_PUB_SECTIONS
9334 True if the @code{.debug_pubtypes} and @code{.debug_pubnames} sections should be emitted. These sections are not used on most platforms, and in particular GDB does not use them.
9337 @defmac ASM_OUTPUT_DWARF_DELTA (@var{stream}, @var{size}, @var{label1}, @var{label2})
9338 A C statement to issue assembly directives that create a difference
9339 @var{lab1} minus @var{lab2}, using an integer of the given @var{size}.
9342 @defmac ASM_OUTPUT_DWARF_VMS_DELTA (@var{stream}, @var{size}, @var{label1}, @var{label2})
9343 A C statement to issue assembly directives that create a difference
9344 between the two given labels in system defined units, e.g. instruction
9345 slots on IA64 VMS, using an integer of the given size.
9348 @defmac ASM_OUTPUT_DWARF_OFFSET (@var{stream}, @var{size}, @var{label}, @var{section})
9349 A C statement to issue assembly directives that create a
9350 section-relative reference to the given @var{label}, using an integer of the
9351 given @var{size}. The label is known to be defined in the given @var{section}.
9354 @defmac ASM_OUTPUT_DWARF_PCREL (@var{stream}, @var{size}, @var{label})
9355 A C statement to issue assembly directives that create a self-relative
9356 reference to the given @var{label}, using an integer of the given @var{size}.
9359 @defmac ASM_OUTPUT_DWARF_TABLE_REF (@var{label})
9360 A C statement to issue assembly directives that create a reference to
9361 the DWARF table identifier @var{label} from the current section. This
9362 is used on some systems to avoid garbage collecting a DWARF table which
9363 is referenced by a function.
9366 @deftypefn {Target Hook} void TARGET_ASM_OUTPUT_DWARF_DTPREL (FILE *@var{file}, int @var{size}, rtx @var{x})
9367 If defined, this target hook is a function which outputs a DTP-relative
9368 reference to the given TLS symbol of the specified size.
9371 @defmac PUT_SDB_@dots{}
9372 Define these macros to override the assembler syntax for the special
9373 SDB assembler directives. See @file{sdbout.c} for a list of these
9374 macros and their arguments. If the standard syntax is used, you need
9375 not define them yourself.
9379 Some assemblers do not support a semicolon as a delimiter, even between
9380 SDB assembler directives. In that case, define this macro to be the
9381 delimiter to use (usually @samp{\n}). It is not necessary to define
9382 a new set of @code{PUT_SDB_@var{op}} macros if this is the only change
9386 @defmac SDB_ALLOW_UNKNOWN_REFERENCES
9387 Define this macro to allow references to unknown structure,
9388 union, or enumeration tags to be emitted. Standard COFF does not
9389 allow handling of unknown references, MIPS ECOFF has support for
9393 @defmac SDB_ALLOW_FORWARD_REFERENCES
9394 Define this macro to allow references to structure, union, or
9395 enumeration tags that have not yet been seen to be handled. Some
9396 assemblers choke if forward tags are used, while some require it.
9399 @defmac SDB_OUTPUT_SOURCE_LINE (@var{stream}, @var{line})
9400 A C statement to output SDB debugging information before code for line
9401 number @var{line} of the current source file to the stdio stream
9402 @var{stream}. The default is to emit an @code{.ln} directive.
9407 @subsection Macros for VMS Debug Format
9409 @c prevent bad page break with this line
9410 Here are macros for VMS debug format.
9412 @defmac VMS_DEBUGGING_INFO
9413 Define this macro if GCC should produce debugging output for VMS
9414 in response to the @option{-g} option. The default behavior for VMS
9415 is to generate minimal debug info for a traceback in the absence of
9416 @option{-g} unless explicitly overridden with @option{-g0}. This
9417 behavior is controlled by @code{OPTIMIZATION_OPTIONS} and
9418 @code{TARGET_OPTION_OVERRIDE}.
9421 @node Floating Point
9422 @section Cross Compilation and Floating Point
9423 @cindex cross compilation and floating point
9424 @cindex floating point and cross compilation
9426 While all modern machines use twos-complement representation for integers,
9427 there are a variety of representations for floating point numbers. This
9428 means that in a cross-compiler the representation of floating point numbers
9429 in the compiled program may be different from that used in the machine
9430 doing the compilation.
9432 Because different representation systems may offer different amounts of
9433 range and precision, all floating point constants must be represented in
9434 the target machine's format. Therefore, the cross compiler cannot
9435 safely use the host machine's floating point arithmetic; it must emulate
9436 the target's arithmetic. To ensure consistency, GCC always uses
9437 emulation to work with floating point values, even when the host and
9438 target floating point formats are identical.
9440 The following macros are provided by @file{real.h} for the compiler to
9441 use. All parts of the compiler which generate or optimize
9442 floating-point calculations must use these macros. They may evaluate
9443 their operands more than once, so operands must not have side effects.
9445 @defmac REAL_VALUE_TYPE
9446 The C data type to be used to hold a floating point value in the target
9447 machine's format. Typically this is a @code{struct} containing an
9448 array of @code{HOST_WIDE_INT}, but all code should treat it as an opaque
9452 @deftypefn Macro int REAL_VALUES_EQUAL (REAL_VALUE_TYPE @var{x}, REAL_VALUE_TYPE @var{y})
9453 Compares for equality the two values, @var{x} and @var{y}. If the target
9454 floating point format supports negative zeroes and/or NaNs,
9455 @samp{REAL_VALUES_EQUAL (-0.0, 0.0)} is true, and
9456 @samp{REAL_VALUES_EQUAL (NaN, NaN)} is false.
9459 @deftypefn Macro int REAL_VALUES_LESS (REAL_VALUE_TYPE @var{x}, REAL_VALUE_TYPE @var{y})
9460 Tests whether @var{x} is less than @var{y}.
9463 @deftypefn Macro HOST_WIDE_INT REAL_VALUE_FIX (REAL_VALUE_TYPE @var{x})
9464 Truncates @var{x} to a signed integer, rounding toward zero.
9467 @deftypefn Macro {unsigned HOST_WIDE_INT} REAL_VALUE_UNSIGNED_FIX (REAL_VALUE_TYPE @var{x})
9468 Truncates @var{x} to an unsigned integer, rounding toward zero. If
9469 @var{x} is negative, returns zero.
9472 @deftypefn Macro REAL_VALUE_TYPE REAL_VALUE_ATOF (const char *@var{string}, enum machine_mode @var{mode})
9473 Converts @var{string} into a floating point number in the target machine's
9474 representation for mode @var{mode}. This routine can handle both
9475 decimal and hexadecimal floating point constants, using the syntax
9476 defined by the C language for both.
9479 @deftypefn Macro int REAL_VALUE_NEGATIVE (REAL_VALUE_TYPE @var{x})
9480 Returns 1 if @var{x} is negative (including negative zero), 0 otherwise.
9483 @deftypefn Macro int REAL_VALUE_ISINF (REAL_VALUE_TYPE @var{x})
9484 Determines whether @var{x} represents infinity (positive or negative).
9487 @deftypefn Macro int REAL_VALUE_ISNAN (REAL_VALUE_TYPE @var{x})
9488 Determines whether @var{x} represents a ``NaN'' (not-a-number).
9491 @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})
9492 Calculates an arithmetic operation on the two floating point values
9493 @var{x} and @var{y}, storing the result in @var{output} (which must be a
9496 The operation to be performed is specified by @var{code}. Only the
9497 following codes are supported: @code{PLUS_EXPR}, @code{MINUS_EXPR},
9498 @code{MULT_EXPR}, @code{RDIV_EXPR}, @code{MAX_EXPR}, @code{MIN_EXPR}.
9500 If @code{REAL_ARITHMETIC} is asked to evaluate division by zero and the
9501 target's floating point format cannot represent infinity, it will call
9502 @code{abort}. Callers should check for this situation first, using
9503 @code{MODE_HAS_INFINITIES}. @xref{Storage Layout}.
9506 @deftypefn Macro REAL_VALUE_TYPE REAL_VALUE_NEGATE (REAL_VALUE_TYPE @var{x})
9507 Returns the negative of the floating point value @var{x}.
9510 @deftypefn Macro REAL_VALUE_TYPE REAL_VALUE_ABS (REAL_VALUE_TYPE @var{x})
9511 Returns the absolute value of @var{x}.
9514 @deftypefn Macro REAL_VALUE_TYPE REAL_VALUE_TRUNCATE (REAL_VALUE_TYPE @var{mode}, enum machine_mode @var{x})
9515 Truncates the floating point value @var{x} to fit in @var{mode}. The
9516 return value is still a full-size @code{REAL_VALUE_TYPE}, but it has an
9517 appropriate bit pattern to be output as a floating constant whose
9518 precision accords with mode @var{mode}.
9521 @deftypefn Macro void REAL_VALUE_TO_INT (HOST_WIDE_INT @var{low}, HOST_WIDE_INT @var{high}, REAL_VALUE_TYPE @var{x})
9522 Converts a floating point value @var{x} into a double-precision integer
9523 which is then stored into @var{low} and @var{high}. If the value is not
9524 integral, it is truncated.
9527 @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})
9528 Converts a double-precision integer found in @var{low} and @var{high},
9529 into a floating point value which is then stored into @var{x}. The
9530 value is truncated to fit in mode @var{mode}.
9533 @node Mode Switching
9534 @section Mode Switching Instructions
9535 @cindex mode switching
9536 The following macros control mode switching optimizations:
9538 @defmac OPTIMIZE_MODE_SWITCHING (@var{entity})
9539 Define this macro if the port needs extra instructions inserted for mode
9540 switching in an optimizing compilation.
9542 For an example, the SH4 can perform both single and double precision
9543 floating point operations, but to perform a single precision operation,
9544 the FPSCR PR bit has to be cleared, while for a double precision
9545 operation, this bit has to be set. Changing the PR bit requires a general
9546 purpose register as a scratch register, hence these FPSCR sets have to
9547 be inserted before reload, i.e.@: you can't put this into instruction emitting
9548 or @code{TARGET_MACHINE_DEPENDENT_REORG}.
9550 You can have multiple entities that are mode-switched, and select at run time
9551 which entities actually need it. @code{OPTIMIZE_MODE_SWITCHING} should
9552 return nonzero for any @var{entity} that needs mode-switching.
9553 If you define this macro, you also have to define
9554 @code{NUM_MODES_FOR_MODE_SWITCHING}, @code{MODE_NEEDED},
9555 @code{MODE_PRIORITY_TO_MODE} and @code{EMIT_MODE_SET}.
9556 @code{MODE_AFTER}, @code{MODE_ENTRY}, and @code{MODE_EXIT}
9560 @defmac NUM_MODES_FOR_MODE_SWITCHING
9561 If you define @code{OPTIMIZE_MODE_SWITCHING}, you have to define this as
9562 initializer for an array of integers. Each initializer element
9563 N refers to an entity that needs mode switching, and specifies the number
9564 of different modes that might need to be set for this entity.
9565 The position of the initializer in the initializer---starting counting at
9566 zero---determines the integer that is used to refer to the mode-switched
9568 In macros that take mode arguments / yield a mode result, modes are
9569 represented as numbers 0 @dots{} N @minus{} 1. N is used to specify that no mode
9570 switch is needed / supplied.
9573 @defmac MODE_NEEDED (@var{entity}, @var{insn})
9574 @var{entity} is an integer specifying a mode-switched entity. If
9575 @code{OPTIMIZE_MODE_SWITCHING} is defined, you must define this macro to
9576 return an integer value not larger than the corresponding element in
9577 @code{NUM_MODES_FOR_MODE_SWITCHING}, to denote the mode that @var{entity} must
9578 be switched into prior to the execution of @var{insn}.
9581 @defmac MODE_AFTER (@var{mode}, @var{insn})
9582 If this macro is defined, it is evaluated for every @var{insn} during
9583 mode switching. It determines the mode that an insn results in (if
9584 different from the incoming mode).
9587 @defmac MODE_ENTRY (@var{entity})
9588 If this macro is defined, it is evaluated for every @var{entity} that needs
9589 mode switching. It should evaluate to an integer, which is a mode that
9590 @var{entity} is assumed to be switched to at function entry. If @code{MODE_ENTRY}
9591 is defined then @code{MODE_EXIT} must be defined.
9594 @defmac MODE_EXIT (@var{entity})
9595 If this macro is defined, it is evaluated for every @var{entity} that needs
9596 mode switching. It should evaluate to an integer, which is a mode that
9597 @var{entity} is assumed to be switched to at function exit. If @code{MODE_EXIT}
9598 is defined then @code{MODE_ENTRY} must be defined.
9601 @defmac MODE_PRIORITY_TO_MODE (@var{entity}, @var{n})
9602 This macro specifies the order in which modes for @var{entity} are processed.
9603 0 is the highest priority, @code{NUM_MODES_FOR_MODE_SWITCHING[@var{entity}] - 1} the
9604 lowest. The value of the macro should be an integer designating a mode
9605 for @var{entity}. For any fixed @var{entity}, @code{mode_priority_to_mode}
9606 (@var{entity}, @var{n}) shall be a bijection in 0 @dots{}
9607 @code{num_modes_for_mode_switching[@var{entity}] - 1}.
9610 @defmac EMIT_MODE_SET (@var{entity}, @var{mode}, @var{hard_regs_live})
9611 Generate one or more insns to set @var{entity} to @var{mode}.
9612 @var{hard_reg_live} is the set of hard registers live at the point where
9613 the insn(s) are to be inserted.
9616 @node Target Attributes
9617 @section Defining target-specific uses of @code{__attribute__}
9618 @cindex target attributes
9619 @cindex machine attributes
9620 @cindex attributes, target-specific
9622 Target-specific attributes may be defined for functions, data and types.
9623 These are described using the following target hooks; they also need to
9624 be documented in @file{extend.texi}.
9626 @deftypevr {Target Hook} {const struct attribute_spec *} TARGET_ATTRIBUTE_TABLE
9627 If defined, this target hook points to an array of @samp{struct
9628 attribute_spec} (defined in @file{tree.h}) specifying the machine
9629 specific attributes for this target and some of the restrictions on the
9630 entities to which these attributes are applied and the arguments they
9634 @deftypefn {Target Hook} bool TARGET_ATTRIBUTE_TAKES_IDENTIFIER_P (const_tree @var{name})
9635 If defined, this target hook is a function which returns true if the
9636 machine-specific attribute named @var{name} expects an identifier
9637 given as its first argument to be passed on as a plain identifier, not
9638 subjected to name lookup. If this is not defined, the default is
9639 false for all machine-specific attributes.
9642 @deftypefn {Target Hook} int TARGET_COMP_TYPE_ATTRIBUTES (const_tree @var{type1}, const_tree @var{type2})
9643 If defined, this target hook is a function which returns zero if the attributes on
9644 @var{type1} and @var{type2} are incompatible, one if they are compatible,
9645 and two if they are nearly compatible (which causes a warning to be
9646 generated). If this is not defined, machine-specific attributes are
9647 supposed always to be compatible.
9650 @deftypefn {Target Hook} void TARGET_SET_DEFAULT_TYPE_ATTRIBUTES (tree @var{type})
9651 If defined, this target hook is a function which assigns default attributes to
9652 the newly defined @var{type}.
9655 @deftypefn {Target Hook} tree TARGET_MERGE_TYPE_ATTRIBUTES (tree @var{type1}, tree @var{type2})
9656 Define this target hook if the merging of type attributes needs special
9657 handling. If defined, the result is a list of the combined
9658 @code{TYPE_ATTRIBUTES} of @var{type1} and @var{type2}. It is assumed
9659 that @code{comptypes} has already been called and returned 1. This
9660 function may call @code{merge_attributes} to handle machine-independent
9664 @deftypefn {Target Hook} tree TARGET_MERGE_DECL_ATTRIBUTES (tree @var{olddecl}, tree @var{newdecl})
9665 Define this target hook if the merging of decl attributes needs special
9666 handling. If defined, the result is a list of the combined
9667 @code{DECL_ATTRIBUTES} of @var{olddecl} and @var{newdecl}.
9668 @var{newdecl} is a duplicate declaration of @var{olddecl}. Examples of
9669 when this is needed are when one attribute overrides another, or when an
9670 attribute is nullified by a subsequent definition. This function may
9671 call @code{merge_attributes} to handle machine-independent merging.
9673 @findex TARGET_DLLIMPORT_DECL_ATTRIBUTES
9674 If the only target-specific handling you require is @samp{dllimport}
9675 for Microsoft Windows targets, you should define the macro
9676 @code{TARGET_DLLIMPORT_DECL_ATTRIBUTES} to @code{1}. The compiler
9677 will then define a function called
9678 @code{merge_dllimport_decl_attributes} which can then be defined as
9679 the expansion of @code{TARGET_MERGE_DECL_ATTRIBUTES}. You can also
9680 add @code{handle_dll_attribute} in the attribute table for your port
9681 to perform initial processing of the @samp{dllimport} and
9682 @samp{dllexport} attributes. This is done in @file{i386/cygwin.h} and
9683 @file{i386/i386.c}, for example.
9686 @deftypefn {Target Hook} bool TARGET_VALID_DLLIMPORT_ATTRIBUTE_P (const_tree @var{decl})
9687 @var{decl} is a variable or function with @code{__attribute__((dllimport))} specified. Use this hook if the target needs to add extra validation checks to @code{handle_dll_attribute}.
9690 @defmac TARGET_DECLSPEC
9691 Define this macro to a nonzero value if you want to treat
9692 @code{__declspec(X)} as equivalent to @code{__attribute((X))}. By
9693 default, this behavior is enabled only for targets that define
9694 @code{TARGET_DLLIMPORT_DECL_ATTRIBUTES}. The current implementation
9695 of @code{__declspec} is via a built-in macro, but you should not rely
9696 on this implementation detail.
9699 @deftypefn {Target Hook} void TARGET_INSERT_ATTRIBUTES (tree @var{node}, tree *@var{attr_ptr})
9700 Define this target hook if you want to be able to add attributes to a decl
9701 when it is being created. This is normally useful for back ends which
9702 wish to implement a pragma by using the attributes which correspond to
9703 the pragma's effect. The @var{node} argument is the decl which is being
9704 created. The @var{attr_ptr} argument is a pointer to the attribute list
9705 for this decl. The list itself should not be modified, since it may be
9706 shared with other decls, but attributes may be chained on the head of
9707 the list and @code{*@var{attr_ptr}} modified to point to the new
9708 attributes, or a copy of the list may be made if further changes are
9712 @deftypefn {Target Hook} bool TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P (const_tree @var{fndecl})
9714 This target hook returns @code{true} if it is ok to inline @var{fndecl}
9715 into the current function, despite its having target-specific
9716 attributes, @code{false} otherwise. By default, if a function has a
9717 target specific attribute attached to it, it will not be inlined.
9720 @deftypefn {Target Hook} bool TARGET_OPTION_VALID_ATTRIBUTE_P (tree @var{fndecl}, tree @var{name}, tree @var{args}, int @var{flags})
9721 This hook is called to parse the @code{attribute(option("..."))}, and
9722 it allows the function to set different target machine compile time
9723 options for the current function that might be different than the
9724 options specified on the command line. The hook should return
9725 @code{true} if the options are valid.
9727 The hook should set the @var{DECL_FUNCTION_SPECIFIC_TARGET} field in
9728 the function declaration to hold a pointer to a target specific
9729 @var{struct cl_target_option} structure.
9732 @deftypefn {Target Hook} void TARGET_OPTION_SAVE (struct cl_target_option *@var{ptr})
9733 This hook is called to save any additional target specific information
9734 in the @var{struct cl_target_option} structure for function specific
9736 @xref{Option file format}.
9739 @deftypefn {Target Hook} void TARGET_OPTION_RESTORE (struct cl_target_option *@var{ptr})
9740 This hook is called to restore any additional target specific
9741 information in the @var{struct cl_target_option} structure for
9742 function specific options.
9745 @deftypefn {Target Hook} void TARGET_OPTION_PRINT (FILE *@var{file}, int @var{indent}, struct cl_target_option *@var{ptr})
9746 This hook is called to print any additional target specific
9747 information in the @var{struct cl_target_option} structure for
9748 function specific options.
9751 @deftypefn {Target Hook} bool TARGET_OPTION_PRAGMA_PARSE (tree @var{args}, tree @var{pop_target})
9752 This target hook parses the options for @code{#pragma GCC option} to
9753 set the machine specific options for functions that occur later in the
9754 input stream. The options should be the same as handled by the
9755 @code{TARGET_OPTION_VALID_ATTRIBUTE_P} hook.
9758 @deftypefn {Target Hook} void TARGET_OPTION_OVERRIDE (void)
9759 Sometimes certain combinations of command options do not make sense on
9760 a particular target machine. You can override the hook
9761 @code{TARGET_OPTION_OVERRIDE} to take account of this. This hooks is called
9762 once just after all the command options have been parsed.
9764 Don't use this hook to turn on various extra optimizations for
9765 @option{-O}. That is what @code{OPTIMIZATION_OPTIONS} is for.
9767 If you need to do something whenever the optimization level is
9768 changed via the optimize attribute or pragma, see
9769 @code{TARGET_OVERRIDE_OPTIONS_AFTER_CHANGE}
9772 @deftypefn {Target Hook} bool TARGET_CAN_INLINE_P (tree @var{caller}, tree @var{callee})
9773 This target hook returns @code{false} if the @var{caller} function
9774 cannot inline @var{callee}, based on target specific information. By
9775 default, inlining is not allowed if the callee function has function
9776 specific target options and the caller does not use the same options.
9780 @section Emulating TLS
9781 @cindex Emulated TLS
9783 For targets whose psABI does not provide Thread Local Storage via
9784 specific relocations and instruction sequences, an emulation layer is
9785 used. A set of target hooks allows this emulation layer to be
9786 configured for the requirements of a particular target. For instance
9787 the psABI may in fact specify TLS support in terms of an emulation
9790 The emulation layer works by creating a control object for every TLS
9791 object. To access the TLS object, a lookup function is provided
9792 which, when given the address of the control object, will return the
9793 address of the current thread's instance of the TLS object.
9795 @deftypevr {Target Hook} {const char *} TARGET_EMUTLS_GET_ADDRESS
9796 Contains the name of the helper function that uses a TLS control
9797 object to locate a TLS instance. The default causes libgcc's
9798 emulated TLS helper function to be used.
9801 @deftypevr {Target Hook} {const char *} TARGET_EMUTLS_REGISTER_COMMON
9802 Contains the name of the helper function that should be used at
9803 program startup to register TLS objects that are implicitly
9804 initialized to zero. If this is @code{NULL}, all TLS objects will
9805 have explicit initializers. The default causes libgcc's emulated TLS
9806 registration function to be used.
9809 @deftypevr {Target Hook} {const char *} TARGET_EMUTLS_VAR_SECTION
9810 Contains the name of the section in which TLS control variables should
9811 be placed. The default of @code{NULL} allows these to be placed in
9815 @deftypevr {Target Hook} {const char *} TARGET_EMUTLS_TMPL_SECTION
9816 Contains the name of the section in which TLS initializers should be
9817 placed. The default of @code{NULL} allows these to be placed in any
9821 @deftypevr {Target Hook} {const char *} TARGET_EMUTLS_VAR_PREFIX
9822 Contains the prefix to be prepended to TLS control variable names.
9823 The default of @code{NULL} uses a target-specific prefix.
9826 @deftypevr {Target Hook} {const char *} TARGET_EMUTLS_TMPL_PREFIX
9827 Contains the prefix to be prepended to TLS initializer objects. The
9828 default of @code{NULL} uses a target-specific prefix.
9831 @deftypefn {Target Hook} tree TARGET_EMUTLS_VAR_FIELDS (tree @var{type}, tree *@var{name})
9832 Specifies a function that generates the FIELD_DECLs for a TLS control
9833 object type. @var{type} is the RECORD_TYPE the fields are for and
9834 @var{name} should be filled with the structure tag, if the default of
9835 @code{__emutls_object} is unsuitable. The default creates a type suitable
9836 for libgcc's emulated TLS function.
9839 @deftypefn {Target Hook} tree TARGET_EMUTLS_VAR_INIT (tree @var{var}, tree @var{decl}, tree @var{tmpl_addr})
9840 Specifies a function that generates the CONSTRUCTOR to initialize a
9841 TLS control object. @var{var} is the TLS control object, @var{decl}
9842 is the TLS object and @var{tmpl_addr} is the address of the
9843 initializer. The default initializes libgcc's emulated TLS control object.
9846 @deftypevr {Target Hook} bool TARGET_EMUTLS_VAR_ALIGN_FIXED
9847 Specifies whether the alignment of TLS control variable objects is
9848 fixed and should not be increased as some backends may do to optimize
9849 single objects. The default is false.
9852 @deftypevr {Target Hook} bool TARGET_EMUTLS_DEBUG_FORM_TLS_ADDRESS
9853 Specifies whether a DWARF @code{DW_OP_form_tls_address} location descriptor
9854 may be used to describe emulated TLS control objects.
9857 @node MIPS Coprocessors
9858 @section Defining coprocessor specifics for MIPS targets.
9859 @cindex MIPS coprocessor-definition macros
9861 The MIPS specification allows MIPS implementations to have as many as 4
9862 coprocessors, each with as many as 32 private registers. GCC supports
9863 accessing these registers and transferring values between the registers
9864 and memory using asm-ized variables. For example:
9867 register unsigned int cp0count asm ("c0r1");
9873 (``c0r1'' is the default name of register 1 in coprocessor 0; alternate
9874 names may be added as described below, or the default names may be
9875 overridden entirely in @code{SUBTARGET_CONDITIONAL_REGISTER_USAGE}.)
9877 Coprocessor registers are assumed to be epilogue-used; sets to them will
9878 be preserved even if it does not appear that the register is used again
9879 later in the function.
9881 Another note: according to the MIPS spec, coprocessor 1 (if present) is
9882 the FPU@. One accesses COP1 registers through standard mips
9883 floating-point support; they are not included in this mechanism.
9885 There is one macro used in defining the MIPS coprocessor interface which
9886 you may want to override in subtargets; it is described below.
9888 @defmac ALL_COP_ADDITIONAL_REGISTER_NAMES
9889 A comma-separated list (with leading comma) of pairs describing the
9890 alternate names of coprocessor registers. The format of each entry should be
9892 @{ @var{alternatename}, @var{register_number}@}
9898 @section Parameters for Precompiled Header Validity Checking
9899 @cindex parameters, precompiled headers
9901 @deftypefn {Target Hook} {void *} TARGET_GET_PCH_VALIDITY (size_t *@var{sz})
9902 This hook returns a pointer to the data needed by
9903 @code{TARGET_PCH_VALID_P} and sets
9904 @samp{*@var{sz}} to the size of the data in bytes.
9907 @deftypefn {Target Hook} {const char *} TARGET_PCH_VALID_P (const void *@var{data}, size_t @var{sz})
9908 This hook checks whether the options used to create a PCH file are
9909 compatible with the current settings. It returns @code{NULL}
9910 if so and a suitable error message if not. Error messages will
9911 be presented to the user and must be localized using @samp{_(@var{msg})}.
9913 @var{data} is the data that was returned by @code{TARGET_GET_PCH_VALIDITY}
9914 when the PCH file was created and @var{sz} is the size of that data in bytes.
9915 It's safe to assume that the data was created by the same version of the
9916 compiler, so no format checking is needed.
9918 The default definition of @code{default_pch_valid_p} should be
9919 suitable for most targets.
9922 @deftypefn {Target Hook} {const char *} TARGET_CHECK_PCH_TARGET_FLAGS (int @var{pch_flags})
9923 If this hook is nonnull, the default implementation of
9924 @code{TARGET_PCH_VALID_P} will use it to check for compatible values
9925 of @code{target_flags}. @var{pch_flags} specifies the value that
9926 @code{target_flags} had when the PCH file was created. The return
9927 value is the same as for @code{TARGET_PCH_VALID_P}.
9931 @section C++ ABI parameters
9932 @cindex parameters, c++ abi
9934 @deftypefn {Target Hook} tree TARGET_CXX_GUARD_TYPE (void)
9935 Define this hook to override the integer type used for guard variables.
9936 These are used to implement one-time construction of static objects. The
9937 default is long_long_integer_type_node.
9940 @deftypefn {Target Hook} bool TARGET_CXX_GUARD_MASK_BIT (void)
9941 This hook determines how guard variables are used. It should return
9942 @code{false} (the default) if the first byte should be used. A return value of
9943 @code{true} indicates that only the least significant bit should be used.
9946 @deftypefn {Target Hook} tree TARGET_CXX_GET_COOKIE_SIZE (tree @var{type})
9947 This hook returns the size of the cookie to use when allocating an array
9948 whose elements have the indicated @var{type}. Assumes that it is already
9949 known that a cookie is needed. The default is
9950 @code{max(sizeof (size_t), alignof(type))}, as defined in section 2.7 of the
9951 IA64/Generic C++ ABI@.
9954 @deftypefn {Target Hook} bool TARGET_CXX_COOKIE_HAS_SIZE (void)
9955 This hook should return @code{true} if the element size should be stored in
9956 array cookies. The default is to return @code{false}.
9959 @deftypefn {Target Hook} int TARGET_CXX_IMPORT_EXPORT_CLASS (tree @var{type}, int @var{import_export})
9960 If defined by a backend this hook allows the decision made to export
9961 class @var{type} to be overruled. Upon entry @var{import_export}
9962 will contain 1 if the class is going to be exported, @minus{}1 if it is going
9963 to be imported and 0 otherwise. This function should return the
9964 modified value and perform any other actions necessary to support the
9965 backend's targeted operating system.
9968 @deftypefn {Target Hook} bool TARGET_CXX_CDTOR_RETURNS_THIS (void)
9969 This hook should return @code{true} if constructors and destructors return
9970 the address of the object created/destroyed. The default is to return
9974 @deftypefn {Target Hook} bool TARGET_CXX_KEY_METHOD_MAY_BE_INLINE (void)
9975 This hook returns true if the key method for a class (i.e., the method
9976 which, if defined in the current translation unit, causes the virtual
9977 table to be emitted) may be an inline function. Under the standard
9978 Itanium C++ ABI the key method may be an inline function so long as
9979 the function is not declared inline in the class definition. Under
9980 some variants of the ABI, an inline function can never be the key
9981 method. The default is to return @code{true}.
9984 @deftypefn {Target Hook} void TARGET_CXX_DETERMINE_CLASS_DATA_VISIBILITY (tree @var{decl})
9985 @var{decl} is a virtual table, virtual table table, typeinfo object, or other similar implicit class data object that will be emitted with external linkage in this translation unit. No ELF visibility has been explicitly specified. If the target needs to specify a visibility other than that of the containing class, use this hook to set @code{DECL_VISIBILITY} and @code{DECL_VISIBILITY_SPECIFIED}.
9988 @deftypefn {Target Hook} bool TARGET_CXX_CLASS_DATA_ALWAYS_COMDAT (void)
9989 This hook returns true (the default) if virtual tables and other
9990 similar implicit class data objects are always COMDAT if they have
9991 external linkage. If this hook returns false, then class data for
9992 classes whose virtual table will be emitted in only one translation
9993 unit will not be COMDAT.
9996 @deftypefn {Target Hook} bool TARGET_CXX_LIBRARY_RTTI_COMDAT (void)
9997 This hook returns true (the default) if the RTTI information for
9998 the basic types which is defined in the C++ runtime should always
9999 be COMDAT, false if it should not be COMDAT.
10002 @deftypefn {Target Hook} bool TARGET_CXX_USE_AEABI_ATEXIT (void)
10003 This hook returns true if @code{__aeabi_atexit} (as defined by the ARM EABI)
10004 should be used to register static destructors when @option{-fuse-cxa-atexit}
10005 is in effect. The default is to return false to use @code{__cxa_atexit}.
10008 @deftypefn {Target Hook} bool TARGET_CXX_USE_ATEXIT_FOR_CXA_ATEXIT (void)
10009 This hook returns true if the target @code{atexit} function can be used
10010 in the same manner as @code{__cxa_atexit} to register C++ static
10011 destructors. This requires that @code{atexit}-registered functions in
10012 shared libraries are run in the correct order when the libraries are
10013 unloaded. The default is to return false.
10016 @deftypefn {Target Hook} void TARGET_CXX_ADJUST_CLASS_AT_DEFINITION (tree @var{type})
10017 @var{type} is a C++ class (i.e., RECORD_TYPE or UNION_TYPE) that has just been defined. Use this hook to make adjustments to the class (eg, tweak visibility or perform any other required target modifications).
10020 @node Named Address Spaces
10021 @section Adding support for named address spaces
10022 @cindex named address spaces
10024 The draft technical report of the ISO/IEC JTC1 S22 WG14 N1275
10025 standards committee, @cite{Programming Languages - C - Extensions to
10026 support embedded processors}, specifies a syntax for embedded
10027 processors to specify alternate address spaces. You can configure a
10028 GCC port to support section 5.1 of the draft report to add support for
10029 address spaces other than the default address space. These address
10030 spaces are new keywords that are similar to the @code{volatile} and
10031 @code{const} type attributes.
10033 Pointers to named address spaces can have a different size than
10034 pointers to the generic address space.
10036 For example, the SPU port uses the @code{__ea} address space to refer
10037 to memory in the host processor, rather than memory local to the SPU
10038 processor. Access to memory in the @code{__ea} address space involves
10039 issuing DMA operations to move data between the host processor and the
10040 local processor memory address space. Pointers in the @code{__ea}
10041 address space are either 32 bits or 64 bits based on the
10042 @option{-mea32} or @option{-mea64} switches (native SPU pointers are
10045 Internally, address spaces are represented as a small integer in the
10046 range 0 to 15 with address space 0 being reserved for the generic
10049 To register a named address space qualifier keyword with the C front end,
10050 the target may call the @code{c_register_addr_space} routine. For example,
10051 the SPU port uses the following to declare @code{__ea} as the keyword for
10052 named address space #1:
10054 #define ADDR_SPACE_EA 1
10055 c_register_addr_space ("__ea", ADDR_SPACE_EA);
10058 @deftypefn {Target Hook} {enum machine_mode} TARGET_ADDR_SPACE_POINTER_MODE (addr_space_t @var{address_space})
10059 Define this to return the machine mode to use for pointers to
10060 @var{address_space} if the target supports named address spaces.
10061 The default version of this hook returns @code{ptr_mode} for the
10062 generic address space only.
10065 @deftypefn {Target Hook} {enum machine_mode} TARGET_ADDR_SPACE_ADDRESS_MODE (addr_space_t @var{address_space})
10066 Define this to return the machine mode to use for addresses in
10067 @var{address_space} if the target supports named address spaces.
10068 The default version of this hook returns @code{Pmode} for the
10069 generic address space only.
10072 @deftypefn {Target Hook} bool TARGET_ADDR_SPACE_VALID_POINTER_MODE (enum machine_mode @var{mode}, addr_space_t @var{as})
10073 Define this to return nonzero if the port can handle pointers
10074 with machine mode @var{mode} to address space @var{as}. This target
10075 hook is the same as the @code{TARGET_VALID_POINTER_MODE} target hook,
10076 except that it includes explicit named address space support. The default
10077 version of this hook returns true for the modes returned by either the
10078 @code{TARGET_ADDR_SPACE_POINTER_MODE} or @code{TARGET_ADDR_SPACE_ADDRESS_MODE}
10079 target hooks for the given address space.
10082 @deftypefn {Target Hook} bool TARGET_ADDR_SPACE_LEGITIMATE_ADDRESS_P (enum machine_mode @var{mode}, rtx @var{exp}, bool @var{strict}, addr_space_t @var{as})
10083 Define this to return true if @var{exp} is a valid address for mode
10084 @var{mode} in the named address space @var{as}. The @var{strict}
10085 parameter says whether strict addressing is in effect after reload has
10086 finished. This target hook is the same as the
10087 @code{TARGET_LEGITIMATE_ADDRESS_P} target hook, except that it includes
10088 explicit named address space support.
10091 @deftypefn {Target Hook} rtx TARGET_ADDR_SPACE_LEGITIMIZE_ADDRESS (rtx @var{x}, rtx @var{oldx}, enum machine_mode @var{mode}, addr_space_t @var{as})
10092 Define this to modify an invalid address @var{x} to be a valid address
10093 with mode @var{mode} in the named address space @var{as}. This target
10094 hook is the same as the @code{TARGET_LEGITIMIZE_ADDRESS} target hook,
10095 except that it includes explicit named address space support.
10098 @deftypefn {Target Hook} bool TARGET_ADDR_SPACE_SUBSET_P (addr_space_t @var{superset}, addr_space_t @var{subset})
10099 Define this to return whether the @var{subset} named address space is
10100 contained within the @var{superset} named address space. Pointers to
10101 a named address space that is a subset of another named address space
10102 will be converted automatically without a cast if used together in
10103 arithmetic operations. Pointers to a superset address space can be
10104 converted to pointers to a subset address space via explicit casts.
10107 @deftypefn {Target Hook} rtx TARGET_ADDR_SPACE_CONVERT (rtx @var{op}, tree @var{from_type}, tree @var{to_type})
10108 Define this to convert the pointer expression represented by the RTL
10109 @var{op} with type @var{from_type} that points to a named address
10110 space to a new pointer expression with type @var{to_type} that points
10111 to a different named address space. When this hook it called, it is
10112 guaranteed that one of the two address spaces is a subset of the other,
10113 as determined by the @code{TARGET_ADDR_SPACE_SUBSET_P} target hook.
10117 @section Miscellaneous Parameters
10118 @cindex parameters, miscellaneous
10120 @c prevent bad page break with this line
10121 Here are several miscellaneous parameters.
10123 @defmac HAS_LONG_COND_BRANCH
10124 Define this boolean macro to indicate whether or not your architecture
10125 has conditional branches that can span all of memory. It is used in
10126 conjunction with an optimization that partitions hot and cold basic
10127 blocks into separate sections of the executable. If this macro is
10128 set to false, gcc will convert any conditional branches that attempt
10129 to cross between sections into unconditional branches or indirect jumps.
10132 @defmac HAS_LONG_UNCOND_BRANCH
10133 Define this boolean macro to indicate whether or not your architecture
10134 has unconditional branches that can span all of memory. It is used in
10135 conjunction with an optimization that partitions hot and cold basic
10136 blocks into separate sections of the executable. If this macro is
10137 set to false, gcc will convert any unconditional branches that attempt
10138 to cross between sections into indirect jumps.
10141 @defmac CASE_VECTOR_MODE
10142 An alias for a machine mode name. This is the machine mode that
10143 elements of a jump-table should have.
10146 @defmac CASE_VECTOR_SHORTEN_MODE (@var{min_offset}, @var{max_offset}, @var{body})
10147 Optional: return the preferred mode for an @code{addr_diff_vec}
10148 when the minimum and maximum offset are known. If you define this,
10149 it enables extra code in branch shortening to deal with @code{addr_diff_vec}.
10150 To make this work, you also have to define @code{INSN_ALIGN} and
10151 make the alignment for @code{addr_diff_vec} explicit.
10152 The @var{body} argument is provided so that the offset_unsigned and scale
10153 flags can be updated.
10156 @defmac CASE_VECTOR_PC_RELATIVE
10157 Define this macro to be a C expression to indicate when jump-tables
10158 should contain relative addresses. You need not define this macro if
10159 jump-tables never contain relative addresses, or jump-tables should
10160 contain relative addresses only when @option{-fPIC} or @option{-fPIC}
10164 @deftypefn {Target Hook} {unsigned int} TARGET_CASE_VALUES_THRESHOLD (void)
10165 This function return the smallest number of different values for which it
10166 is best to use a jump-table instead of a tree of conditional branches.
10167 The default is four for machines with a @code{casesi} instruction and
10168 five otherwise. This is best for most machines.
10171 @defmac CASE_USE_BIT_TESTS
10172 Define this macro to be a C expression to indicate whether C switch
10173 statements may be implemented by a sequence of bit tests. This is
10174 advantageous on processors that can efficiently implement left shift
10175 of 1 by the number of bits held in a register, but inappropriate on
10176 targets that would require a loop. By default, this macro returns
10177 @code{true} if the target defines an @code{ashlsi3} pattern, and
10178 @code{false} otherwise.
10181 @defmac WORD_REGISTER_OPERATIONS
10182 Define this macro if operations between registers with integral mode
10183 smaller than a word are always performed on the entire register.
10184 Most RISC machines have this property and most CISC machines do not.
10187 @defmac LOAD_EXTEND_OP (@var{mem_mode})
10188 Define this macro to be a C expression indicating when insns that read
10189 memory in @var{mem_mode}, an integral mode narrower than a word, set the
10190 bits outside of @var{mem_mode} to be either the sign-extension or the
10191 zero-extension of the data read. Return @code{SIGN_EXTEND} for values
10192 of @var{mem_mode} for which the
10193 insn sign-extends, @code{ZERO_EXTEND} for which it zero-extends, and
10194 @code{UNKNOWN} for other modes.
10196 This macro is not called with @var{mem_mode} non-integral or with a width
10197 greater than or equal to @code{BITS_PER_WORD}, so you may return any
10198 value in this case. Do not define this macro if it would always return
10199 @code{UNKNOWN}. On machines where this macro is defined, you will normally
10200 define it as the constant @code{SIGN_EXTEND} or @code{ZERO_EXTEND}.
10202 You may return a non-@code{UNKNOWN} value even if for some hard registers
10203 the sign extension is not performed, if for the @code{REGNO_REG_CLASS}
10204 of these hard registers @code{CANNOT_CHANGE_MODE_CLASS} returns nonzero
10205 when the @var{from} mode is @var{mem_mode} and the @var{to} mode is any
10206 integral mode larger than this but not larger than @code{word_mode}.
10208 You must return @code{UNKNOWN} if for some hard registers that allow this
10209 mode, @code{CANNOT_CHANGE_MODE_CLASS} says that they cannot change to
10210 @code{word_mode}, but that they can change to another integral mode that
10211 is larger then @var{mem_mode} but still smaller than @code{word_mode}.
10214 @defmac SHORT_IMMEDIATES_SIGN_EXTEND
10215 Define this macro if loading short immediate values into registers sign
10219 @defmac FIXUNS_TRUNC_LIKE_FIX_TRUNC
10220 Define this macro if the same instructions that convert a floating
10221 point number to a signed fixed point number also convert validly to an
10225 @deftypefn {Target Hook} {unsigned int} TARGET_MIN_DIVISIONS_FOR_RECIP_MUL (enum machine_mode @var{mode})
10226 When @option{-ffast-math} is in effect, GCC tries to optimize
10227 divisions by the same divisor, by turning them into multiplications by
10228 the reciprocal. This target hook specifies the minimum number of divisions
10229 that should be there for GCC to perform the optimization for a variable
10230 of mode @var{mode}. The default implementation returns 3 if the machine
10231 has an instruction for the division, and 2 if it does not.
10235 The maximum number of bytes that a single instruction can move quickly
10236 between memory and registers or between two memory locations.
10239 @defmac MAX_MOVE_MAX
10240 The maximum number of bytes that a single instruction can move quickly
10241 between memory and registers or between two memory locations. If this
10242 is undefined, the default is @code{MOVE_MAX}. Otherwise, it is the
10243 constant value that is the largest value that @code{MOVE_MAX} can have
10247 @defmac SHIFT_COUNT_TRUNCATED
10248 A C expression that is nonzero if on this machine the number of bits
10249 actually used for the count of a shift operation is equal to the number
10250 of bits needed to represent the size of the object being shifted. When
10251 this macro is nonzero, the compiler will assume that it is safe to omit
10252 a sign-extend, zero-extend, and certain bitwise `and' instructions that
10253 truncates the count of a shift operation. On machines that have
10254 instructions that act on bit-fields at variable positions, which may
10255 include `bit test' instructions, a nonzero @code{SHIFT_COUNT_TRUNCATED}
10256 also enables deletion of truncations of the values that serve as
10257 arguments to bit-field instructions.
10259 If both types of instructions truncate the count (for shifts) and
10260 position (for bit-field operations), or if no variable-position bit-field
10261 instructions exist, you should define this macro.
10263 However, on some machines, such as the 80386 and the 680x0, truncation
10264 only applies to shift operations and not the (real or pretended)
10265 bit-field operations. Define @code{SHIFT_COUNT_TRUNCATED} to be zero on
10266 such machines. Instead, add patterns to the @file{md} file that include
10267 the implied truncation of the shift instructions.
10269 You need not define this macro if it would always have the value of zero.
10272 @anchor{TARGET_SHIFT_TRUNCATION_MASK}
10273 @deftypefn {Target Hook} {unsigned HOST_WIDE_INT} TARGET_SHIFT_TRUNCATION_MASK (enum machine_mode @var{mode})
10274 This function describes how the standard shift patterns for @var{mode}
10275 deal with shifts by negative amounts or by more than the width of the mode.
10276 @xref{shift patterns}.
10278 On many machines, the shift patterns will apply a mask @var{m} to the
10279 shift count, meaning that a fixed-width shift of @var{x} by @var{y} is
10280 equivalent to an arbitrary-width shift of @var{x} by @var{y & m}. If
10281 this is true for mode @var{mode}, the function should return @var{m},
10282 otherwise it should return 0. A return value of 0 indicates that no
10283 particular behavior is guaranteed.
10285 Note that, unlike @code{SHIFT_COUNT_TRUNCATED}, this function does
10286 @emph{not} apply to general shift rtxes; it applies only to instructions
10287 that are generated by the named shift patterns.
10289 The default implementation of this function returns
10290 @code{GET_MODE_BITSIZE (@var{mode}) - 1} if @code{SHIFT_COUNT_TRUNCATED}
10291 and 0 otherwise. This definition is always safe, but if
10292 @code{SHIFT_COUNT_TRUNCATED} is false, and some shift patterns
10293 nevertheless truncate the shift count, you may get better code
10297 @defmac TRULY_NOOP_TRUNCATION (@var{outprec}, @var{inprec})
10298 A C expression which is nonzero if on this machine it is safe to
10299 ``convert'' an integer of @var{inprec} bits to one of @var{outprec}
10300 bits (where @var{outprec} is smaller than @var{inprec}) by merely
10301 operating on it as if it had only @var{outprec} bits.
10303 On many machines, this expression can be 1.
10305 @c rearranged this, removed the phrase "it is reported that". this was
10306 @c to fix an overfull hbox. --mew 10feb93
10307 When @code{TRULY_NOOP_TRUNCATION} returns 1 for a pair of sizes for
10308 modes for which @code{MODES_TIEABLE_P} is 0, suboptimal code can result.
10309 If this is the case, making @code{TRULY_NOOP_TRUNCATION} return 0 in
10310 such cases may improve things.
10313 @deftypefn {Target Hook} int TARGET_MODE_REP_EXTENDED (enum machine_mode @var{mode}, enum machine_mode @var{rep_mode})
10314 The representation of an integral mode can be such that the values
10315 are always extended to a wider integral mode. Return
10316 @code{SIGN_EXTEND} if values of @var{mode} are represented in
10317 sign-extended form to @var{rep_mode}. Return @code{UNKNOWN}
10318 otherwise. (Currently, none of the targets use zero-extended
10319 representation this way so unlike @code{LOAD_EXTEND_OP},
10320 @code{TARGET_MODE_REP_EXTENDED} is expected to return either
10321 @code{SIGN_EXTEND} or @code{UNKNOWN}. Also no target extends
10322 @var{mode} to @var{rep_mode} so that @var{rep_mode} is not the next
10323 widest integral mode and currently we take advantage of this fact.)
10325 Similarly to @code{LOAD_EXTEND_OP} you may return a non-@code{UNKNOWN}
10326 value even if the extension is not performed on certain hard registers
10327 as long as for the @code{REGNO_REG_CLASS} of these hard registers
10328 @code{CANNOT_CHANGE_MODE_CLASS} returns nonzero.
10330 Note that @code{TARGET_MODE_REP_EXTENDED} and @code{LOAD_EXTEND_OP}
10331 describe two related properties. If you define
10332 @code{TARGET_MODE_REP_EXTENDED (mode, word_mode)} you probably also want
10333 to define @code{LOAD_EXTEND_OP (mode)} to return the same type of
10336 In order to enforce the representation of @code{mode},
10337 @code{TRULY_NOOP_TRUNCATION} should return false when truncating to
10341 @defmac STORE_FLAG_VALUE
10342 A C expression describing the value returned by a comparison operator
10343 with an integral mode and stored by a store-flag instruction
10344 (@samp{cstore@var{mode}4}) when the condition is true. This description must
10345 apply to @emph{all} the @samp{cstore@var{mode}4} patterns and all the
10346 comparison operators whose results have a @code{MODE_INT} mode.
10348 A value of 1 or @minus{}1 means that the instruction implementing the
10349 comparison operator returns exactly 1 or @minus{}1 when the comparison is true
10350 and 0 when the comparison is false. Otherwise, the value indicates
10351 which bits of the result are guaranteed to be 1 when the comparison is
10352 true. This value is interpreted in the mode of the comparison
10353 operation, which is given by the mode of the first operand in the
10354 @samp{cstore@var{mode}4} pattern. Either the low bit or the sign bit of
10355 @code{STORE_FLAG_VALUE} be on. Presently, only those bits are used by
10358 If @code{STORE_FLAG_VALUE} is neither 1 or @minus{}1, the compiler will
10359 generate code that depends only on the specified bits. It can also
10360 replace comparison operators with equivalent operations if they cause
10361 the required bits to be set, even if the remaining bits are undefined.
10362 For example, on a machine whose comparison operators return an
10363 @code{SImode} value and where @code{STORE_FLAG_VALUE} is defined as
10364 @samp{0x80000000}, saying that just the sign bit is relevant, the
10368 (ne:SI (and:SI @var{x} (const_int @var{power-of-2})) (const_int 0))
10372 can be converted to
10375 (ashift:SI @var{x} (const_int @var{n}))
10379 where @var{n} is the appropriate shift count to move the bit being
10380 tested into the sign bit.
10382 There is no way to describe a machine that always sets the low-order bit
10383 for a true value, but does not guarantee the value of any other bits,
10384 but we do not know of any machine that has such an instruction. If you
10385 are trying to port GCC to such a machine, include an instruction to
10386 perform a logical-and of the result with 1 in the pattern for the
10387 comparison operators and let us know at @email{gcc@@gcc.gnu.org}.
10389 Often, a machine will have multiple instructions that obtain a value
10390 from a comparison (or the condition codes). Here are rules to guide the
10391 choice of value for @code{STORE_FLAG_VALUE}, and hence the instructions
10396 Use the shortest sequence that yields a valid definition for
10397 @code{STORE_FLAG_VALUE}. It is more efficient for the compiler to
10398 ``normalize'' the value (convert it to, e.g., 1 or 0) than for the
10399 comparison operators to do so because there may be opportunities to
10400 combine the normalization with other operations.
10403 For equal-length sequences, use a value of 1 or @minus{}1, with @minus{}1 being
10404 slightly preferred on machines with expensive jumps and 1 preferred on
10408 As a second choice, choose a value of @samp{0x80000001} if instructions
10409 exist that set both the sign and low-order bits but do not define the
10413 Otherwise, use a value of @samp{0x80000000}.
10416 Many machines can produce both the value chosen for
10417 @code{STORE_FLAG_VALUE} and its negation in the same number of
10418 instructions. On those machines, you should also define a pattern for
10419 those cases, e.g., one matching
10422 (set @var{A} (neg:@var{m} (ne:@var{m} @var{B} @var{C})))
10425 Some machines can also perform @code{and} or @code{plus} operations on
10426 condition code values with less instructions than the corresponding
10427 @samp{cstore@var{mode}4} insn followed by @code{and} or @code{plus}. On those
10428 machines, define the appropriate patterns. Use the names @code{incscc}
10429 and @code{decscc}, respectively, for the patterns which perform
10430 @code{plus} or @code{minus} operations on condition code values. See
10431 @file{rs6000.md} for some examples. The GNU Superoptizer can be used to
10432 find such instruction sequences on other machines.
10434 If this macro is not defined, the default value, 1, is used. You need
10435 not define @code{STORE_FLAG_VALUE} if the machine has no store-flag
10436 instructions, or if the value generated by these instructions is 1.
10439 @defmac FLOAT_STORE_FLAG_VALUE (@var{mode})
10440 A C expression that gives a nonzero @code{REAL_VALUE_TYPE} value that is
10441 returned when comparison operators with floating-point results are true.
10442 Define this macro on machines that have comparison operations that return
10443 floating-point values. If there are no such operations, do not define
10447 @defmac VECTOR_STORE_FLAG_VALUE (@var{mode})
10448 A C expression that gives a rtx representing the nonzero true element
10449 for vector comparisons. The returned rtx should be valid for the inner
10450 mode of @var{mode} which is guaranteed to be a vector mode. Define
10451 this macro on machines that have vector comparison operations that
10452 return a vector result. If there are no such operations, do not define
10453 this macro. Typically, this macro is defined as @code{const1_rtx} or
10454 @code{constm1_rtx}. This macro may return @code{NULL_RTX} to prevent
10455 the compiler optimizing such vector comparison operations for the
10459 @defmac CLZ_DEFINED_VALUE_AT_ZERO (@var{mode}, @var{value})
10460 @defmacx CTZ_DEFINED_VALUE_AT_ZERO (@var{mode}, @var{value})
10461 A C expression that indicates whether the architecture defines a value
10462 for @code{clz} or @code{ctz} with a zero operand.
10463 A result of @code{0} indicates the value is undefined.
10464 If the value is defined for only the RTL expression, the macro should
10465 evaluate to @code{1}; if the value applies also to the corresponding optab
10466 entry (which is normally the case if it expands directly into
10467 the corresponding RTL), then the macro should evaluate to @code{2}.
10468 In the cases where the value is defined, @var{value} should be set to
10471 If this macro is not defined, the value of @code{clz} or
10472 @code{ctz} at zero is assumed to be undefined.
10474 This macro must be defined if the target's expansion for @code{ffs}
10475 relies on a particular value to get correct results. Otherwise it
10476 is not necessary, though it may be used to optimize some corner cases, and
10477 to provide a default expansion for the @code{ffs} optab.
10479 Note that regardless of this macro the ``definedness'' of @code{clz}
10480 and @code{ctz} at zero do @emph{not} extend to the builtin functions
10481 visible to the user. Thus one may be free to adjust the value at will
10482 to match the target expansion of these operations without fear of
10487 An alias for the machine mode for pointers. On most machines, define
10488 this to be the integer mode corresponding to the width of a hardware
10489 pointer; @code{SImode} on 32-bit machine or @code{DImode} on 64-bit machines.
10490 On some machines you must define this to be one of the partial integer
10491 modes, such as @code{PSImode}.
10493 The width of @code{Pmode} must be at least as large as the value of
10494 @code{POINTER_SIZE}. If it is not equal, you must define the macro
10495 @code{POINTERS_EXTEND_UNSIGNED} to specify how pointers are extended
10499 @defmac FUNCTION_MODE
10500 An alias for the machine mode used for memory references to functions
10501 being called, in @code{call} RTL expressions. On most CISC machines,
10502 where an instruction can begin at any byte address, this should be
10503 @code{QImode}. On most RISC machines, where all instructions have fixed
10504 size and alignment, this should be a mode with the same size and alignment
10505 as the machine instruction words - typically @code{SImode} or @code{HImode}.
10508 @defmac STDC_0_IN_SYSTEM_HEADERS
10509 In normal operation, the preprocessor expands @code{__STDC__} to the
10510 constant 1, to signify that GCC conforms to ISO Standard C@. On some
10511 hosts, like Solaris, the system compiler uses a different convention,
10512 where @code{__STDC__} is normally 0, but is 1 if the user specifies
10513 strict conformance to the C Standard.
10515 Defining @code{STDC_0_IN_SYSTEM_HEADERS} makes GNU CPP follows the host
10516 convention when processing system header files, but when processing user
10517 files @code{__STDC__} will always expand to 1.
10520 @defmac NO_IMPLICIT_EXTERN_C
10521 Define this macro if the system header files support C++ as well as C@.
10522 This macro inhibits the usual method of using system header files in
10523 C++, which is to pretend that the file's contents are enclosed in
10524 @samp{extern "C" @{@dots{}@}}.
10529 @defmac REGISTER_TARGET_PRAGMAS ()
10530 Define this macro if you want to implement any target-specific pragmas.
10531 If defined, it is a C expression which makes a series of calls to
10532 @code{c_register_pragma} or @code{c_register_pragma_with_expansion}
10533 for each pragma. The macro may also do any
10534 setup required for the pragmas.
10536 The primary reason to define this macro is to provide compatibility with
10537 other compilers for the same target. In general, we discourage
10538 definition of target-specific pragmas for GCC@.
10540 If the pragma can be implemented by attributes then you should consider
10541 defining the target hook @samp{TARGET_INSERT_ATTRIBUTES} as well.
10543 Preprocessor macros that appear on pragma lines are not expanded. All
10544 @samp{#pragma} directives that do not match any registered pragma are
10545 silently ignored, unless the user specifies @option{-Wunknown-pragmas}.
10548 @deftypefun void c_register_pragma (const char *@var{space}, const char *@var{name}, void (*@var{callback}) (struct cpp_reader *))
10549 @deftypefunx void c_register_pragma_with_expansion (const char *@var{space}, const char *@var{name}, void (*@var{callback}) (struct cpp_reader *))
10551 Each call to @code{c_register_pragma} or
10552 @code{c_register_pragma_with_expansion} establishes one pragma. The
10553 @var{callback} routine will be called when the preprocessor encounters a
10557 #pragma [@var{space}] @var{name} @dots{}
10560 @var{space} is the case-sensitive namespace of the pragma, or
10561 @code{NULL} to put the pragma in the global namespace. The callback
10562 routine receives @var{pfile} as its first argument, which can be passed
10563 on to cpplib's functions if necessary. You can lex tokens after the
10564 @var{name} by calling @code{pragma_lex}. Tokens that are not read by the
10565 callback will be silently ignored. The end of the line is indicated by
10566 a token of type @code{CPP_EOF}. Macro expansion occurs on the
10567 arguments of pragmas registered with
10568 @code{c_register_pragma_with_expansion} but not on the arguments of
10569 pragmas registered with @code{c_register_pragma}.
10571 Note that the use of @code{pragma_lex} is specific to the C and C++
10572 compilers. It will not work in the Java or Fortran compilers, or any
10573 other language compilers for that matter. Thus if @code{pragma_lex} is going
10574 to be called from target-specific code, it must only be done so when
10575 building the C and C++ compilers. This can be done by defining the
10576 variables @code{c_target_objs} and @code{cxx_target_objs} in the
10577 target entry in the @file{config.gcc} file. These variables should name
10578 the target-specific, language-specific object file which contains the
10579 code that uses @code{pragma_lex}. Note it will also be necessary to add a
10580 rule to the makefile fragment pointed to by @code{tmake_file} that shows
10581 how to build this object file.
10586 @defmac HANDLE_SYSV_PRAGMA
10587 Define this macro (to a value of 1) if you want the System V style
10588 pragmas @samp{#pragma pack(<n>)} and @samp{#pragma weak <name>
10589 [=<value>]} to be supported by gcc.
10591 The pack pragma specifies the maximum alignment (in bytes) of fields
10592 within a structure, in much the same way as the @samp{__aligned__} and
10593 @samp{__packed__} @code{__attribute__}s do. A pack value of zero resets
10594 the behavior to the default.
10596 A subtlety for Microsoft Visual C/C++ style bit-field packing
10597 (e.g.@: -mms-bitfields) for targets that support it:
10598 When a bit-field is inserted into a packed record, the whole size
10599 of the underlying type is used by one or more same-size adjacent
10600 bit-fields (that is, if its long:3, 32 bits is used in the record,
10601 and any additional adjacent long bit-fields are packed into the same
10602 chunk of 32 bits. However, if the size changes, a new field of that
10603 size is allocated).
10605 If both MS bit-fields and @samp{__attribute__((packed))} are used,
10606 the latter will take precedence. If @samp{__attribute__((packed))} is
10607 used on a single field when MS bit-fields are in use, it will take
10608 precedence for that field, but the alignment of the rest of the structure
10609 may affect its placement.
10611 The weak pragma only works if @code{SUPPORTS_WEAK} and
10612 @code{ASM_WEAKEN_LABEL} are defined. If enabled it allows the creation
10613 of specifically named weak labels, optionally with a value.
10618 @defmac HANDLE_PRAGMA_PACK_PUSH_POP
10619 Define this macro (to a value of 1) if you want to support the Win32
10620 style pragmas @samp{#pragma pack(push[,@var{n}])} and @samp{#pragma
10621 pack(pop)}. The @samp{pack(push,[@var{n}])} pragma specifies the maximum
10622 alignment (in bytes) of fields within a structure, in much the same way as
10623 the @samp{__aligned__} and @samp{__packed__} @code{__attribute__}s do. A
10624 pack value of zero resets the behavior to the default. Successive
10625 invocations of this pragma cause the previous values to be stacked, so
10626 that invocations of @samp{#pragma pack(pop)} will return to the previous
10630 @defmac HANDLE_PRAGMA_PACK_WITH_EXPANSION
10631 Define this macro, as well as
10632 @code{HANDLE_SYSV_PRAGMA}, if macros should be expanded in the
10633 arguments of @samp{#pragma pack}.
10636 @deftypevr {Target Hook} bool TARGET_HANDLE_PRAGMA_EXTERN_PREFIX
10637 True if @code{#pragma extern_prefix} is to be supported.
10640 @defmac TARGET_DEFAULT_PACK_STRUCT
10641 If your target requires a structure packing default other than 0 (meaning
10642 the machine default), define this macro to the necessary value (in bytes).
10643 This must be a value that would also be valid to use with
10644 @samp{#pragma pack()} (that is, a small power of two).
10647 @defmac DOLLARS_IN_IDENTIFIERS
10648 Define this macro to control use of the character @samp{$} in
10649 identifier names for the C family of languages. 0 means @samp{$} is
10650 not allowed by default; 1 means it is allowed. 1 is the default;
10651 there is no need to define this macro in that case.
10654 @defmac NO_DOLLAR_IN_LABEL
10655 Define this macro if the assembler does not accept the character
10656 @samp{$} in label names. By default constructors and destructors in
10657 G++ have @samp{$} in the identifiers. If this macro is defined,
10658 @samp{.} is used instead.
10661 @defmac NO_DOT_IN_LABEL
10662 Define this macro if the assembler does not accept the character
10663 @samp{.} in label names. By default constructors and destructors in G++
10664 have names that use @samp{.}. If this macro is defined, these names
10665 are rewritten to avoid @samp{.}.
10668 @defmac INSN_SETS_ARE_DELAYED (@var{insn})
10669 Define this macro as a C expression that is nonzero if it is safe for the
10670 delay slot scheduler to place instructions in the delay slot of @var{insn},
10671 even if they appear to use a resource set or clobbered in @var{insn}.
10672 @var{insn} is always a @code{jump_insn} or an @code{insn}; GCC knows that
10673 every @code{call_insn} has this behavior. On machines where some @code{insn}
10674 or @code{jump_insn} is really a function call and hence has this behavior,
10675 you should define this macro.
10677 You need not define this macro if it would always return zero.
10680 @defmac INSN_REFERENCES_ARE_DELAYED (@var{insn})
10681 Define this macro as a C expression that is nonzero if it is safe for the
10682 delay slot scheduler to place instructions in the delay slot of @var{insn},
10683 even if they appear to set or clobber a resource referenced in @var{insn}.
10684 @var{insn} is always a @code{jump_insn} or an @code{insn}. On machines where
10685 some @code{insn} or @code{jump_insn} is really a function call and its operands
10686 are registers whose use is actually in the subroutine it calls, you should
10687 define this macro. Doing so allows the delay slot scheduler to move
10688 instructions which copy arguments into the argument registers into the delay
10689 slot of @var{insn}.
10691 You need not define this macro if it would always return zero.
10694 @defmac MULTIPLE_SYMBOL_SPACES
10695 Define this macro as a C expression that is nonzero if, in some cases,
10696 global symbols from one translation unit may not be bound to undefined
10697 symbols in another translation unit without user intervention. For
10698 instance, under Microsoft Windows symbols must be explicitly imported
10699 from shared libraries (DLLs).
10701 You need not define this macro if it would always evaluate to zero.
10704 @deftypefn {Target Hook} tree TARGET_MD_ASM_CLOBBERS (tree @var{outputs}, tree @var{inputs}, tree @var{clobbers})
10705 This target hook should add to @var{clobbers} @code{STRING_CST} trees for
10706 any hard regs the port wishes to automatically clobber for an asm.
10707 It should return the result of the last @code{tree_cons} used to add a
10708 clobber. The @var{outputs}, @var{inputs} and @var{clobber} lists are the
10709 corresponding parameters to the asm and may be inspected to avoid
10710 clobbering a register that is an input or output of the asm. You can use
10711 @code{tree_overlaps_hard_reg_set}, declared in @file{tree.h}, to test
10712 for overlap with regards to asm-declared registers.
10715 @defmac MATH_LIBRARY
10716 Define this macro as a C string constant for the linker argument to link
10717 in the system math library, minus the initial @samp{"-l"}, or
10718 @samp{""} if the target does not have a
10719 separate math library.
10721 You need only define this macro if the default of @samp{"m"} is wrong.
10724 @defmac LIBRARY_PATH_ENV
10725 Define this macro as a C string constant for the environment variable that
10726 specifies where the linker should look for libraries.
10728 You need only define this macro if the default of @samp{"LIBRARY_PATH"}
10732 @defmac TARGET_POSIX_IO
10733 Define this macro if the target supports the following POSIX@ file
10734 functions, access, mkdir and file locking with fcntl / F_SETLKW@.
10735 Defining @code{TARGET_POSIX_IO} will enable the test coverage code
10736 to use file locking when exiting a program, which avoids race conditions
10737 if the program has forked. It will also create directories at run-time
10738 for cross-profiling.
10741 @defmac MAX_CONDITIONAL_EXECUTE
10743 A C expression for the maximum number of instructions to execute via
10744 conditional execution instructions instead of a branch. A value of
10745 @code{BRANCH_COST}+1 is the default if the machine does not use cc0, and
10746 1 if it does use cc0.
10749 @defmac IFCVT_MODIFY_TESTS (@var{ce_info}, @var{true_expr}, @var{false_expr})
10750 Used if the target needs to perform machine-dependent modifications on the
10751 conditionals used for turning basic blocks into conditionally executed code.
10752 @var{ce_info} points to a data structure, @code{struct ce_if_block}, which
10753 contains information about the currently processed blocks. @var{true_expr}
10754 and @var{false_expr} are the tests that are used for converting the
10755 then-block and the else-block, respectively. Set either @var{true_expr} or
10756 @var{false_expr} to a null pointer if the tests cannot be converted.
10759 @defmac IFCVT_MODIFY_MULTIPLE_TESTS (@var{ce_info}, @var{bb}, @var{true_expr}, @var{false_expr})
10760 Like @code{IFCVT_MODIFY_TESTS}, but used when converting more complicated
10761 if-statements into conditions combined by @code{and} and @code{or} operations.
10762 @var{bb} contains the basic block that contains the test that is currently
10763 being processed and about to be turned into a condition.
10766 @defmac IFCVT_MODIFY_INSN (@var{ce_info}, @var{pattern}, @var{insn})
10767 A C expression to modify the @var{PATTERN} of an @var{INSN} that is to
10768 be converted to conditional execution format. @var{ce_info} points to
10769 a data structure, @code{struct ce_if_block}, which contains information
10770 about the currently processed blocks.
10773 @defmac IFCVT_MODIFY_FINAL (@var{ce_info})
10774 A C expression to perform any final machine dependent modifications in
10775 converting code to conditional execution. The involved basic blocks
10776 can be found in the @code{struct ce_if_block} structure that is pointed
10777 to by @var{ce_info}.
10780 @defmac IFCVT_MODIFY_CANCEL (@var{ce_info})
10781 A C expression to cancel any machine dependent modifications in
10782 converting code to conditional execution. The involved basic blocks
10783 can be found in the @code{struct ce_if_block} structure that is pointed
10784 to by @var{ce_info}.
10787 @defmac IFCVT_INIT_EXTRA_FIELDS (@var{ce_info})
10788 A C expression to initialize any extra fields in a @code{struct ce_if_block}
10789 structure, which are defined by the @code{IFCVT_EXTRA_FIELDS} macro.
10792 @defmac IFCVT_EXTRA_FIELDS
10793 If defined, it should expand to a set of field declarations that will be
10794 added to the @code{struct ce_if_block} structure. These should be initialized
10795 by the @code{IFCVT_INIT_EXTRA_FIELDS} macro.
10798 @deftypefn {Target Hook} void TARGET_MACHINE_DEPENDENT_REORG (void)
10799 If non-null, this hook performs a target-specific pass over the
10800 instruction stream. The compiler will run it at all optimization levels,
10801 just before the point at which it normally does delayed-branch scheduling.
10803 The exact purpose of the hook varies from target to target. Some use
10804 it to do transformations that are necessary for correctness, such as
10805 laying out in-function constant pools or avoiding hardware hazards.
10806 Others use it as an opportunity to do some machine-dependent optimizations.
10808 You need not implement the hook if it has nothing to do. The default
10809 definition is null.
10812 @deftypefn {Target Hook} void TARGET_INIT_BUILTINS (void)
10813 Define this hook if you have any machine-specific built-in functions
10814 that need to be defined. It should be a function that performs the
10817 Machine specific built-in functions can be useful to expand special machine
10818 instructions that would otherwise not normally be generated because
10819 they have no equivalent in the source language (for example, SIMD vector
10820 instructions or prefetch instructions).
10822 To create a built-in function, call the function
10823 @code{lang_hooks.builtin_function}
10824 which is defined by the language front end. You can use any type nodes set
10825 up by @code{build_common_tree_nodes} and @code{build_common_tree_nodes_2};
10826 only language front ends that use those two functions will call
10827 @samp{TARGET_INIT_BUILTINS}.
10830 @deftypefn {Target Hook} tree TARGET_BUILTIN_DECL (unsigned @var{code}, bool @var{initialize_p})
10831 Define this hook if you have any machine-specific built-in functions
10832 that need to be defined. It should be a function that returns the
10833 builtin function declaration for the builtin function code @var{code}.
10834 If there is no such builtin and it cannot be initialized at this time
10835 if @var{initialize_p} is true the function should return @code{NULL_TREE}.
10836 If @var{code} is out of range the function should return
10837 @code{error_mark_node}.
10840 @deftypefn {Target Hook} rtx TARGET_EXPAND_BUILTIN (tree @var{exp}, rtx @var{target}, rtx @var{subtarget}, enum machine_mode @var{mode}, int @var{ignore})
10842 Expand a call to a machine specific built-in function that was set up by
10843 @samp{TARGET_INIT_BUILTINS}. @var{exp} is the expression for the
10844 function call; the result should go to @var{target} if that is
10845 convenient, and have mode @var{mode} if that is convenient.
10846 @var{subtarget} may be used as the target for computing one of
10847 @var{exp}'s operands. @var{ignore} is nonzero if the value is to be
10848 ignored. This function should return the result of the call to the
10852 @deftypefn {Target Hook} tree TARGET_RESOLVE_OVERLOADED_BUILTIN (unsigned int @var{loc}, tree @var{fndecl}, void *@var{arglist})
10853 Select a replacement for a machine specific built-in function that
10854 was set up by @samp{TARGET_INIT_BUILTINS}. This is done
10855 @emph{before} regular type checking, and so allows the target to
10856 implement a crude form of function overloading. @var{fndecl} is the
10857 declaration of the built-in function. @var{arglist} is the list of
10858 arguments passed to the built-in function. The result is a
10859 complete expression that implements the operation, usually
10860 another @code{CALL_EXPR}.
10861 @var{arglist} really has type @samp{VEC(tree,gc)*}
10864 @deftypefn {Target Hook} tree TARGET_FOLD_BUILTIN (tree @var{fndecl}, int @var{n_args}, tree *@var{argp}, bool @var{ignore})
10865 Fold a call to a machine specific built-in function that was set up by
10866 @samp{TARGET_INIT_BUILTINS}. @var{fndecl} is the declaration of the
10867 built-in function. @var{n_args} is the number of arguments passed to
10868 the function; the arguments themselves are pointed to by @var{argp}.
10869 The result is another tree containing a simplified expression for the
10870 call's result. If @var{ignore} is true the value will be ignored.
10873 @deftypefn {Target Hook} {const char *} TARGET_INVALID_WITHIN_DOLOOP (const_rtx @var{insn})
10875 Take an instruction in @var{insn} and return NULL if it is valid within a
10876 low-overhead loop, otherwise return a string explaining why doloop
10877 could not be applied.
10879 Many targets use special registers for low-overhead looping. For any
10880 instruction that clobbers these this function should return a string indicating
10881 the reason why the doloop could not be applied.
10882 By default, the RTL loop optimizer does not use a present doloop pattern for
10883 loops containing function calls or branch on table instructions.
10886 @defmac MD_CAN_REDIRECT_BRANCH (@var{branch1}, @var{branch2})
10888 Take a branch insn in @var{branch1} and another in @var{branch2}.
10889 Return true if redirecting @var{branch1} to the destination of
10890 @var{branch2} is possible.
10892 On some targets, branches may have a limited range. Optimizing the
10893 filling of delay slots can result in branches being redirected, and this
10894 may in turn cause a branch offset to overflow.
10897 @deftypefn {Target Hook} bool TARGET_COMMUTATIVE_P (const_rtx @var{x}, int @var{outer_code})
10898 This target hook returns @code{true} if @var{x} is considered to be commutative.
10899 Usually, this is just COMMUTATIVE_P (@var{x}), but the HP PA doesn't consider
10900 PLUS to be commutative inside a MEM@. @var{outer_code} is the rtx code
10901 of the enclosing rtl, if known, otherwise it is UNKNOWN.
10904 @deftypefn {Target Hook} rtx TARGET_ALLOCATE_INITIAL_VALUE (rtx @var{hard_reg})
10906 When the initial value of a hard register has been copied in a pseudo
10907 register, it is often not necessary to actually allocate another register
10908 to this pseudo register, because the original hard register or a stack slot
10909 it has been saved into can be used. @code{TARGET_ALLOCATE_INITIAL_VALUE}
10910 is called at the start of register allocation once for each hard register
10911 that had its initial value copied by using
10912 @code{get_func_hard_reg_initial_val} or @code{get_hard_reg_initial_val}.
10913 Possible values are @code{NULL_RTX}, if you don't want
10914 to do any special allocation, a @code{REG} rtx---that would typically be
10915 the hard register itself, if it is known not to be clobbered---or a
10917 If you are returning a @code{MEM}, this is only a hint for the allocator;
10918 it might decide to use another register anyways.
10919 You may use @code{current_function_leaf_function} in the hook, functions
10920 that use @code{REG_N_SETS}, to determine if the hard
10921 register in question will not be clobbered.
10922 The default value of this hook is @code{NULL}, which disables any special
10926 @deftypefn {Target Hook} int TARGET_UNSPEC_MAY_TRAP_P (const_rtx @var{x}, unsigned @var{flags})
10927 This target hook returns nonzero if @var{x}, an @code{unspec} or
10928 @code{unspec_volatile} operation, might cause a trap. Targets can use
10929 this hook to enhance precision of analysis for @code{unspec} and
10930 @code{unspec_volatile} operations. You may call @code{may_trap_p_1}
10931 to analyze inner elements of @var{x} in which case @var{flags} should be
10935 @deftypefn {Target Hook} void TARGET_SET_CURRENT_FUNCTION (tree @var{decl})
10936 The compiler invokes this hook whenever it changes its current function
10937 context (@code{cfun}). You can define this function if
10938 the back end needs to perform any initialization or reset actions on a
10939 per-function basis. For example, it may be used to implement function
10940 attributes that affect register usage or code generation patterns.
10941 The argument @var{decl} is the declaration for the new function context,
10942 and may be null to indicate that the compiler has left a function context
10943 and is returning to processing at the top level.
10944 The default hook function does nothing.
10946 GCC sets @code{cfun} to a dummy function context during initialization of
10947 some parts of the back end. The hook function is not invoked in this
10948 situation; you need not worry about the hook being invoked recursively,
10949 or when the back end is in a partially-initialized state.
10950 @code{cfun} might be @code{NULL} to indicate processing at top level,
10951 outside of any function scope.
10954 @defmac TARGET_OBJECT_SUFFIX
10955 Define this macro to be a C string representing the suffix for object
10956 files on your target machine. If you do not define this macro, GCC will
10957 use @samp{.o} as the suffix for object files.
10960 @defmac TARGET_EXECUTABLE_SUFFIX
10961 Define this macro to be a C string representing the suffix to be
10962 automatically added to executable files on your target machine. If you
10963 do not define this macro, GCC will use the null string as the suffix for
10967 @defmac COLLECT_EXPORT_LIST
10968 If defined, @code{collect2} will scan the individual object files
10969 specified on its command line and create an export list for the linker.
10970 Define this macro for systems like AIX, where the linker discards
10971 object files that are not referenced from @code{main} and uses export
10975 @defmac MODIFY_JNI_METHOD_CALL (@var{mdecl})
10976 Define this macro to a C expression representing a variant of the
10977 method call @var{mdecl}, if Java Native Interface (JNI) methods
10978 must be invoked differently from other methods on your target.
10979 For example, on 32-bit Microsoft Windows, JNI methods must be invoked using
10980 the @code{stdcall} calling convention and this macro is then
10981 defined as this expression:
10984 build_type_attribute_variant (@var{mdecl},
10986 (get_identifier ("stdcall"),
10991 @deftypefn {Target Hook} bool TARGET_CANNOT_MODIFY_JUMPS_P (void)
10992 This target hook returns @code{true} past the point in which new jump
10993 instructions could be created. On machines that require a register for
10994 every jump such as the SHmedia ISA of SH5, this point would typically be
10995 reload, so this target hook should be defined to a function such as:
10999 cannot_modify_jumps_past_reload_p ()
11001 return (reload_completed || reload_in_progress);
11006 @deftypefn {Target Hook} reg_class_t TARGET_BRANCH_TARGET_REGISTER_CLASS (void)
11007 This target hook returns a register class for which branch target register
11008 optimizations should be applied. All registers in this class should be
11009 usable interchangeably. After reload, registers in this class will be
11010 re-allocated and loads will be hoisted out of loops and be subjected
11011 to inter-block scheduling.
11014 @deftypefn {Target Hook} bool TARGET_BRANCH_TARGET_REGISTER_CALLEE_SAVED (bool @var{after_prologue_epilogue_gen})
11015 Branch target register optimization will by default exclude callee-saved
11017 that are not already live during the current function; if this target hook
11018 returns true, they will be included. The target code must than make sure
11019 that all target registers in the class returned by
11020 @samp{TARGET_BRANCH_TARGET_REGISTER_CLASS} that might need saving are
11021 saved. @var{after_prologue_epilogue_gen} indicates if prologues and
11022 epilogues have already been generated. Note, even if you only return
11023 true when @var{after_prologue_epilogue_gen} is false, you still are likely
11024 to have to make special provisions in @code{INITIAL_ELIMINATION_OFFSET}
11025 to reserve space for caller-saved target registers.
11028 @deftypefn {Target Hook} bool TARGET_HAVE_CONDITIONAL_EXECUTION (void)
11029 This target hook returns true if the target supports conditional execution.
11030 This target hook is required only when the target has several different
11031 modes and they have different conditional execution capability, such as ARM.
11034 @deftypefn {Target Hook} unsigned TARGET_LOOP_UNROLL_ADJUST (unsigned @var{nunroll}, struct loop *@var{loop})
11035 This target hook returns a new value for the number of times @var{loop}
11036 should be unrolled. The parameter @var{nunroll} is the number of times
11037 the loop is to be unrolled. The parameter @var{loop} is a pointer to
11038 the loop, which is going to be checked for unrolling. This target hook
11039 is required only when the target has special constraints like maximum
11040 number of memory accesses.
11043 @defmac POWI_MAX_MULTS
11044 If defined, this macro is interpreted as a signed integer C expression
11045 that specifies the maximum number of floating point multiplications
11046 that should be emitted when expanding exponentiation by an integer
11047 constant inline. When this value is defined, exponentiation requiring
11048 more than this number of multiplications is implemented by calling the
11049 system library's @code{pow}, @code{powf} or @code{powl} routines.
11050 The default value places no upper bound on the multiplication count.
11053 @deftypefn Macro void TARGET_EXTRA_INCLUDES (const char *@var{sysroot}, const char *@var{iprefix}, int @var{stdinc})
11054 This target hook should register any extra include files for the
11055 target. The parameter @var{stdinc} indicates if normal include files
11056 are present. The parameter @var{sysroot} is the system root directory.
11057 The parameter @var{iprefix} is the prefix for the gcc directory.
11060 @deftypefn Macro void TARGET_EXTRA_PRE_INCLUDES (const char *@var{sysroot}, const char *@var{iprefix}, int @var{stdinc})
11061 This target hook should register any extra include files for the
11062 target before any standard headers. The parameter @var{stdinc}
11063 indicates if normal include files are present. The parameter
11064 @var{sysroot} is the system root directory. The parameter
11065 @var{iprefix} is the prefix for the gcc directory.
11068 @deftypefn Macro void TARGET_OPTF (char *@var{path})
11069 This target hook should register special include paths for the target.
11070 The parameter @var{path} is the include to register. On Darwin
11071 systems, this is used for Framework includes, which have semantics
11072 that are different from @option{-I}.
11075 @defmac bool TARGET_USE_LOCAL_THUNK_ALIAS_P (tree @var{fndecl})
11076 This target macro returns @code{true} if it is safe to use a local alias
11077 for a virtual function @var{fndecl} when constructing thunks,
11078 @code{false} otherwise. By default, the macro returns @code{true} for all
11079 functions, if a target supports aliases (i.e.@: defines
11080 @code{ASM_OUTPUT_DEF}), @code{false} otherwise,
11083 @defmac TARGET_FORMAT_TYPES
11084 If defined, this macro is the name of a global variable containing
11085 target-specific format checking information for the @option{-Wformat}
11086 option. The default is to have no target-specific format checks.
11089 @defmac TARGET_N_FORMAT_TYPES
11090 If defined, this macro is the number of entries in
11091 @code{TARGET_FORMAT_TYPES}.
11094 @defmac TARGET_OVERRIDES_FORMAT_ATTRIBUTES
11095 If defined, this macro is the name of a global variable containing
11096 target-specific format overrides for the @option{-Wformat} option. The
11097 default is to have no target-specific format overrides. If defined,
11098 @code{TARGET_FORMAT_TYPES} must be defined, too.
11101 @defmac TARGET_OVERRIDES_FORMAT_ATTRIBUTES_COUNT
11102 If defined, this macro specifies the number of entries in
11103 @code{TARGET_OVERRIDES_FORMAT_ATTRIBUTES}.
11106 @defmac TARGET_OVERRIDES_FORMAT_INIT
11107 If defined, this macro specifies the optional initialization
11108 routine for target specific customizations of the system printf
11109 and scanf formatter settings.
11112 @deftypevr {Target Hook} bool TARGET_RELAXED_ORDERING
11113 If set to @code{true}, means that the target's memory model does not
11114 guarantee that loads which do not depend on one another will access
11115 main memory in the order of the instruction stream; if ordering is
11116 important, an explicit memory barrier must be used. This is true of
11117 many recent processors which implement a policy of ``relaxed,''
11118 ``weak,'' or ``release'' memory consistency, such as Alpha, PowerPC,
11119 and ia64. The default is @code{false}.
11122 @deftypefn {Target Hook} {const char *} TARGET_INVALID_ARG_FOR_UNPROTOTYPED_FN (const_tree @var{typelist}, const_tree @var{funcdecl}, const_tree @var{val})
11123 If defined, this macro returns the diagnostic message when it is
11124 illegal to pass argument @var{val} to function @var{funcdecl}
11125 with prototype @var{typelist}.
11128 @deftypefn {Target Hook} {const char *} TARGET_INVALID_CONVERSION (const_tree @var{fromtype}, const_tree @var{totype})
11129 If defined, this macro returns the diagnostic message when it is
11130 invalid to convert from @var{fromtype} to @var{totype}, or @code{NULL}
11131 if validity should be determined by the front end.
11134 @deftypefn {Target Hook} {const char *} TARGET_INVALID_UNARY_OP (int @var{op}, const_tree @var{type})
11135 If defined, this macro returns the diagnostic message when it is
11136 invalid to apply operation @var{op} (where unary plus is denoted by
11137 @code{CONVERT_EXPR}) to an operand of type @var{type}, or @code{NULL}
11138 if validity should be determined by the front end.
11141 @deftypefn {Target Hook} {const char *} TARGET_INVALID_BINARY_OP (int @var{op}, const_tree @var{type1}, const_tree @var{type2})
11142 If defined, this macro returns the diagnostic message when it is
11143 invalid to apply operation @var{op} to operands of types @var{type1}
11144 and @var{type2}, or @code{NULL} if validity should be determined by
11148 @deftypefn {Target Hook} {const char *} TARGET_INVALID_PARAMETER_TYPE (const_tree @var{type})
11149 If defined, this macro returns the diagnostic message when it is
11150 invalid for functions to include parameters of type @var{type},
11151 or @code{NULL} if validity should be determined by
11152 the front end. This is currently used only by the C and C++ front ends.
11155 @deftypefn {Target Hook} {const char *} TARGET_INVALID_RETURN_TYPE (const_tree @var{type})
11156 If defined, this macro returns the diagnostic message when it is
11157 invalid for functions to have return type @var{type},
11158 or @code{NULL} if validity should be determined by
11159 the front end. This is currently used only by the C and C++ front ends.
11162 @deftypefn {Target Hook} tree TARGET_PROMOTED_TYPE (const_tree @var{type})
11163 If defined, this target hook returns the type to which values of
11164 @var{type} should be promoted when they appear in expressions,
11165 analogous to the integer promotions, or @code{NULL_TREE} to use the
11166 front end's normal promotion rules. This hook is useful when there are
11167 target-specific types with special promotion rules.
11168 This is currently used only by the C and C++ front ends.
11171 @deftypefn {Target Hook} tree TARGET_CONVERT_TO_TYPE (tree @var{type}, tree @var{expr})
11172 If defined, this hook returns the result of converting @var{expr} to
11173 @var{type}. It should return the converted expression,
11174 or @code{NULL_TREE} to apply the front end's normal conversion rules.
11175 This hook is useful when there are target-specific types with special
11177 This is currently used only by the C and C++ front ends.
11180 @defmac TARGET_USE_JCR_SECTION
11181 This macro determines whether to use the JCR section to register Java
11182 classes. By default, TARGET_USE_JCR_SECTION is defined to 1 if both
11183 SUPPORTS_WEAK and TARGET_HAVE_NAMED_SECTIONS are true, else 0.
11187 This macro determines the size of the objective C jump buffer for the
11188 NeXT runtime. By default, OBJC_JBLEN is defined to an innocuous value.
11191 @defmac LIBGCC2_UNWIND_ATTRIBUTE
11192 Define this macro if any target-specific attributes need to be attached
11193 to the functions in @file{libgcc} that provide low-level support for
11194 call stack unwinding. It is used in declarations in @file{unwind-generic.h}
11195 and the associated definitions of those functions.
11198 @deftypefn {Target Hook} void TARGET_UPDATE_STACK_BOUNDARY (void)
11199 Define this macro to update the current function stack boundary if
11203 @deftypefn {Target Hook} rtx TARGET_GET_DRAP_RTX (void)
11204 This hook should return an rtx for Dynamic Realign Argument Pointer (DRAP) if a
11205 different argument pointer register is needed to access the function's
11206 argument list due to stack realignment. Return @code{NULL} if no DRAP
11210 @deftypefn {Target Hook} bool TARGET_ALLOCATE_STACK_SLOTS_FOR_ARGS (void)
11211 When optimization is disabled, this hook indicates whether or not
11212 arguments should be allocated to stack slots. Normally, GCC allocates
11213 stacks slots for arguments when not optimizing in order to make
11214 debugging easier. However, when a function is declared with
11215 @code{__attribute__((naked))}, there is no stack frame, and the compiler
11216 cannot safely move arguments from the registers in which they are passed
11217 to the stack. Therefore, this hook should return true in general, but
11218 false for naked functions. The default implementation always returns true.
11221 @deftypevr {Target Hook} {unsigned HOST_WIDE_INT} TARGET_CONST_ANCHOR
11222 On some architectures it can take multiple instructions to synthesize
11223 a constant. If there is another constant already in a register that
11224 is close enough in value then it is preferable that the new constant
11225 is computed from this register using immediate addition or
11226 subtraction. We accomplish this through CSE. Besides the value of
11227 the constant we also add a lower and an upper constant anchor to the
11228 available expressions. These are then queried when encountering new
11229 constants. The anchors are computed by rounding the constant up and
11230 down to a multiple of the value of @code{TARGET_CONST_ANCHOR}.
11231 @code{TARGET_CONST_ANCHOR} should be the maximum positive value
11232 accepted by immediate-add plus one. We currently assume that the
11233 value of @code{TARGET_CONST_ANCHOR} is a power of 2. For example, on
11234 MIPS, where add-immediate takes a 16-bit signed value,
11235 @code{TARGET_CONST_ANCHOR} is set to @samp{0x8000}. The default value
11236 is zero, which disables this optimization. @end deftypevr