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, 2011
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.
94 Similarly, there is a @code{targetcm} variable for hooks that are
95 specific to front ends for C-family languages, documented as ``C
96 Target Hook''. This is declared in @file{c-family/c-target.h}, the
97 initializer @code{TARGETCM_INITIALIZER} in
98 @file{c-family/c-target-def.h}. If targets initialize @code{targetcm}
99 themselves, they should set @code{target_has_targetcm=yes} in
100 @file{config.gcc}; otherwise a default definition is used.
102 Similarly, there is a @code{targetm_common} variable for hooks that
103 are shared between the compiler driver and the compilers proper,
104 documented as ``Common Target Hook''. This is declared in
105 @file{common/common-target.h}, the initializer
106 @code{TARGETM_COMMON_INITIALIZER} in
107 @file{common/common-target-def.h}. If targets initialize
108 @code{targetm_common} themselves, they should set
109 @code{target_has_targetm_common=yes} in @file{config.gcc}; otherwise a
110 default definition is used.
113 @section Controlling the Compilation Driver, @file{gcc}
115 @cindex controlling the compilation driver
117 @c prevent bad page break with this line
118 You can control the compilation driver.
120 @defmac DRIVER_SELF_SPECS
121 A list of specs for the driver itself. It should be a suitable
122 initializer for an array of strings, with no surrounding braces.
124 The driver applies these specs to its own command line between loading
125 default @file{specs} files (but not command-line specified ones) and
126 choosing the multilib directory or running any subcommands. It
127 applies them in the order given, so each spec can depend on the
128 options added by earlier ones. It is also possible to remove options
129 using @samp{%<@var{option}} in the usual way.
131 This macro can be useful when a port has several interdependent target
132 options. It provides a way of standardizing the command line so
133 that the other specs are easier to write.
135 Do not define this macro if it does not need to do anything.
138 @defmac OPTION_DEFAULT_SPECS
139 A list of specs used to support configure-time default options (i.e.@:
140 @option{--with} options) in the driver. It should be a suitable initializer
141 for an array of structures, each containing two strings, without the
142 outermost pair of surrounding braces.
144 The first item in the pair is the name of the default. This must match
145 the code in @file{config.gcc} for the target. The second item is a spec
146 to apply if a default with this name was specified. The string
147 @samp{%(VALUE)} in the spec will be replaced by the value of the default
148 everywhere it occurs.
150 The driver will apply these specs to its own command line between loading
151 default @file{specs} files and processing @code{DRIVER_SELF_SPECS}, using
152 the same mechanism as @code{DRIVER_SELF_SPECS}.
154 Do not define this macro if it does not need to do anything.
158 A C string constant that tells the GCC driver program options to
159 pass to CPP@. It can also specify how to translate options you
160 give to GCC into options for GCC to pass to the CPP@.
162 Do not define this macro if it does not need to do anything.
165 @defmac CPLUSPLUS_CPP_SPEC
166 This macro is just like @code{CPP_SPEC}, but is used for C++, rather
167 than C@. If you do not define this macro, then the value of
168 @code{CPP_SPEC} (if any) will be used instead.
172 A C string constant that tells the GCC driver program options to
173 pass to @code{cc1}, @code{cc1plus}, @code{f771}, and the other language
175 It can also specify how to translate options you give to GCC into options
176 for GCC to pass to front ends.
178 Do not define this macro if it does not need to do anything.
182 A C string constant that tells the GCC driver program options to
183 pass to @code{cc1plus}. It can also specify how to translate options you
184 give to GCC into options for GCC to pass to the @code{cc1plus}.
186 Do not define this macro if it does not need to do anything.
187 Note that everything defined in CC1_SPEC is already passed to
188 @code{cc1plus} so there is no need to duplicate the contents of
189 CC1_SPEC in CC1PLUS_SPEC@.
193 A C string constant that tells the GCC driver program options to
194 pass to the assembler. It can also specify how to translate options
195 you give to GCC into options for GCC to pass to the assembler.
196 See the file @file{sun3.h} for an example of this.
198 Do not define this macro if it does not need to do anything.
201 @defmac ASM_FINAL_SPEC
202 A C string constant that tells the GCC driver program how to
203 run any programs which cleanup after the normal assembler.
204 Normally, this is not needed. See the file @file{mips.h} for
207 Do not define this macro if it does not need to do anything.
210 @defmac AS_NEEDS_DASH_FOR_PIPED_INPUT
211 Define this macro, with no value, if the driver should give the assembler
212 an argument consisting of a single dash, @option{-}, to instruct it to
213 read from its standard input (which will be a pipe connected to the
214 output of the compiler proper). This argument is given after any
215 @option{-o} option specifying the name of the output file.
217 If you do not define this macro, the assembler is assumed to read its
218 standard input if given no non-option arguments. If your assembler
219 cannot read standard input at all, use a @samp{%@{pipe:%e@}} construct;
220 see @file{mips.h} for instance.
224 A C string constant that tells the GCC driver program options to
225 pass to the linker. It can also specify how to translate options you
226 give to GCC into options for GCC to pass to the linker.
228 Do not define this macro if it does not need to do anything.
232 Another C string constant used much like @code{LINK_SPEC}. The difference
233 between the two is that @code{LIB_SPEC} is used at the end of the
234 command given to the linker.
236 If this macro is not defined, a default is provided that
237 loads the standard C library from the usual place. See @file{gcc.c}.
241 Another C string constant that tells the GCC driver program
242 how and when to place a reference to @file{libgcc.a} into the
243 linker command line. This constant is placed both before and after
244 the value of @code{LIB_SPEC}.
246 If this macro is not defined, the GCC driver provides a default that
247 passes the string @option{-lgcc} to the linker.
250 @defmac REAL_LIBGCC_SPEC
251 By default, if @code{ENABLE_SHARED_LIBGCC} is defined, the
252 @code{LIBGCC_SPEC} is not directly used by the driver program but is
253 instead modified to refer to different versions of @file{libgcc.a}
254 depending on the values of the command line flags @option{-static},
255 @option{-shared}, @option{-static-libgcc}, and @option{-shared-libgcc}. On
256 targets where these modifications are inappropriate, define
257 @code{REAL_LIBGCC_SPEC} instead. @code{REAL_LIBGCC_SPEC} tells the
258 driver how to place a reference to @file{libgcc} on the link command
259 line, but, unlike @code{LIBGCC_SPEC}, it is used unmodified.
262 @defmac USE_LD_AS_NEEDED
263 A macro that controls the modifications to @code{LIBGCC_SPEC}
264 mentioned in @code{REAL_LIBGCC_SPEC}. If nonzero, a spec will be
265 generated that uses --as-needed and the shared libgcc in place of the
266 static exception handler library, when linking without any of
267 @code{-static}, @code{-static-libgcc}, or @code{-shared-libgcc}.
271 If defined, this C string constant is added to @code{LINK_SPEC}.
272 When @code{USE_LD_AS_NEEDED} is zero or undefined, it also affects
273 the modifications to @code{LIBGCC_SPEC} mentioned in
274 @code{REAL_LIBGCC_SPEC}.
277 @defmac STARTFILE_SPEC
278 Another C string constant used much like @code{LINK_SPEC}. The
279 difference between the two is that @code{STARTFILE_SPEC} is used at
280 the very beginning of the command given to the linker.
282 If this macro is not defined, a default is provided that loads the
283 standard C startup file from the usual place. See @file{gcc.c}.
287 Another C string constant used much like @code{LINK_SPEC}. The
288 difference between the two is that @code{ENDFILE_SPEC} is used at
289 the very end of the command given to the linker.
291 Do not define this macro if it does not need to do anything.
294 @defmac THREAD_MODEL_SPEC
295 GCC @code{-v} will print the thread model GCC was configured to use.
296 However, this doesn't work on platforms that are multilibbed on thread
297 models, such as AIX 4.3. On such platforms, define
298 @code{THREAD_MODEL_SPEC} such that it evaluates to a string without
299 blanks that names one of the recognized thread models. @code{%*}, the
300 default value of this macro, will expand to the value of
301 @code{thread_file} set in @file{config.gcc}.
304 @defmac SYSROOT_SUFFIX_SPEC
305 Define this macro to add a suffix to the target sysroot when GCC is
306 configured with a sysroot. This will cause GCC to search for usr/lib,
307 et al, within sysroot+suffix.
310 @defmac SYSROOT_HEADERS_SUFFIX_SPEC
311 Define this macro to add a headers_suffix to the target sysroot when
312 GCC is configured with a sysroot. This will cause GCC to pass the
313 updated sysroot+headers_suffix to CPP, causing it to search for
314 usr/include, et al, within sysroot+headers_suffix.
318 Define this macro to provide additional specifications to put in the
319 @file{specs} file that can be used in various specifications like
322 The definition should be an initializer for an array of structures,
323 containing a string constant, that defines the specification name, and a
324 string constant that provides the specification.
326 Do not define this macro if it does not need to do anything.
328 @code{EXTRA_SPECS} is useful when an architecture contains several
329 related targets, which have various @code{@dots{}_SPECS} which are similar
330 to each other, and the maintainer would like one central place to keep
333 For example, the PowerPC System V.4 targets use @code{EXTRA_SPECS} to
334 define either @code{_CALL_SYSV} when the System V calling sequence is
335 used or @code{_CALL_AIX} when the older AIX-based calling sequence is
338 The @file{config/rs6000/rs6000.h} target file defines:
341 #define EXTRA_SPECS \
342 @{ "cpp_sysv_default", CPP_SYSV_DEFAULT @},
344 #define CPP_SYS_DEFAULT ""
347 The @file{config/rs6000/sysv.h} target file defines:
351 "%@{posix: -D_POSIX_SOURCE @} \
352 %@{mcall-sysv: -D_CALL_SYSV @} \
353 %@{!mcall-sysv: %(cpp_sysv_default) @} \
354 %@{msoft-float: -D_SOFT_FLOAT@} %@{mcpu=403: -D_SOFT_FLOAT@}"
356 #undef CPP_SYSV_DEFAULT
357 #define CPP_SYSV_DEFAULT "-D_CALL_SYSV"
360 while the @file{config/rs6000/eabiaix.h} target file defines
361 @code{CPP_SYSV_DEFAULT} as:
364 #undef CPP_SYSV_DEFAULT
365 #define CPP_SYSV_DEFAULT "-D_CALL_AIX"
369 @defmac LINK_LIBGCC_SPECIAL_1
370 Define this macro if the driver program should find the library
371 @file{libgcc.a}. If you do not define this macro, the driver program will pass
372 the argument @option{-lgcc} to tell the linker to do the search.
375 @defmac LINK_GCC_C_SEQUENCE_SPEC
376 The sequence in which libgcc and libc are specified to the linker.
377 By default this is @code{%G %L %G}.
380 @defmac LINK_COMMAND_SPEC
381 A C string constant giving the complete command line need to execute the
382 linker. When you do this, you will need to update your port each time a
383 change is made to the link command line within @file{gcc.c}. Therefore,
384 define this macro only if you need to completely redefine the command
385 line for invoking the linker and there is no other way to accomplish
386 the effect you need. Overriding this macro may be avoidable by overriding
387 @code{LINK_GCC_C_SEQUENCE_SPEC} instead.
390 @defmac LINK_ELIMINATE_DUPLICATE_LDIRECTORIES
391 A nonzero value causes @command{collect2} to remove duplicate @option{-L@var{directory}} search
392 directories from linking commands. Do not give it a nonzero value if
393 removing duplicate search directories changes the linker's semantics.
396 @hook TARGET_ALWAYS_STRIP_DOTDOT
398 @defmac MULTILIB_DEFAULTS
399 Define this macro as a C expression for the initializer of an array of
400 string to tell the driver program which options are defaults for this
401 target and thus do not need to be handled specially when using
402 @code{MULTILIB_OPTIONS}.
404 Do not define this macro if @code{MULTILIB_OPTIONS} is not defined in
405 the target makefile fragment or if none of the options listed in
406 @code{MULTILIB_OPTIONS} are set by default.
407 @xref{Target Fragment}.
410 @defmac RELATIVE_PREFIX_NOT_LINKDIR
411 Define this macro to tell @command{gcc} that it should only translate
412 a @option{-B} prefix into a @option{-L} linker option if the prefix
413 indicates an absolute file name.
416 @defmac MD_EXEC_PREFIX
417 If defined, this macro is an additional prefix to try after
418 @code{STANDARD_EXEC_PREFIX}. @code{MD_EXEC_PREFIX} is not searched
419 when the compiler is built as a cross
420 compiler. If you define @code{MD_EXEC_PREFIX}, then be sure to add it
421 to the list of directories used to find the assembler in @file{configure.in}.
424 @defmac STANDARD_STARTFILE_PREFIX
425 Define this macro as a C string constant if you wish to override the
426 standard choice of @code{libdir} as the default prefix to
427 try when searching for startup files such as @file{crt0.o}.
428 @code{STANDARD_STARTFILE_PREFIX} is not searched when the compiler
429 is built as a cross compiler.
432 @defmac STANDARD_STARTFILE_PREFIX_1
433 Define this macro as a C string constant if you wish to override the
434 standard choice of @code{/lib} as a prefix to try after the default prefix
435 when searching for startup files such as @file{crt0.o}.
436 @code{STANDARD_STARTFILE_PREFIX_1} is not searched when the compiler
437 is built as a cross compiler.
440 @defmac STANDARD_STARTFILE_PREFIX_2
441 Define this macro as a C string constant if you wish to override the
442 standard choice of @code{/lib} as yet another prefix to try after the
443 default prefix when searching for startup files such as @file{crt0.o}.
444 @code{STANDARD_STARTFILE_PREFIX_2} is not searched when the compiler
445 is built as a cross compiler.
448 @defmac MD_STARTFILE_PREFIX
449 If defined, this macro supplies an additional prefix to try after the
450 standard prefixes. @code{MD_EXEC_PREFIX} is not searched when the
451 compiler is built as a cross compiler.
454 @defmac MD_STARTFILE_PREFIX_1
455 If defined, this macro supplies yet another prefix to try after the
456 standard prefixes. It is not searched when the compiler is built as a
460 @defmac INIT_ENVIRONMENT
461 Define this macro as a C string constant if you wish to set environment
462 variables for programs called by the driver, such as the assembler and
463 loader. The driver passes the value of this macro to @code{putenv} to
464 initialize the necessary environment variables.
467 @defmac LOCAL_INCLUDE_DIR
468 Define this macro as a C string constant if you wish to override the
469 standard choice of @file{/usr/local/include} as the default prefix to
470 try when searching for local header files. @code{LOCAL_INCLUDE_DIR}
471 comes before @code{NATIVE_SYSTEM_HEADER_DIR} (set in
472 @file{config.gcc}, normally @file{/usr/include}) in the search order.
474 Cross compilers do not search either @file{/usr/local/include} or its
478 @defmac NATIVE_SYSTEM_HEADER_COMPONENT
479 The ``component'' corresponding to @code{NATIVE_SYSTEM_HEADER_DIR}.
480 See @code{INCLUDE_DEFAULTS}, below, for the description of components.
481 If you do not define this macro, no component is used.
484 @defmac INCLUDE_DEFAULTS
485 Define this macro if you wish to override the entire default search path
486 for include files. For a native compiler, the default search path
487 usually consists of @code{GCC_INCLUDE_DIR}, @code{LOCAL_INCLUDE_DIR},
488 @code{GPLUSPLUS_INCLUDE_DIR}, and
489 @code{NATIVE_SYSTEM_HEADER_DIR}. In addition, @code{GPLUSPLUS_INCLUDE_DIR}
490 and @code{GCC_INCLUDE_DIR} are defined automatically by @file{Makefile},
491 and specify private search areas for GCC@. The directory
492 @code{GPLUSPLUS_INCLUDE_DIR} is used only for C++ programs.
494 The definition should be an initializer for an array of structures.
495 Each array element should have four elements: the directory name (a
496 string constant), the component name (also a string constant), a flag
497 for C++-only directories,
498 and a flag showing that the includes in the directory don't need to be
499 wrapped in @code{extern @samp{C}} when compiling C++. Mark the end of
500 the array with a null element.
502 The component name denotes what GNU package the include file is part of,
503 if any, in all uppercase letters. For example, it might be @samp{GCC}
504 or @samp{BINUTILS}. If the package is part of a vendor-supplied
505 operating system, code the component name as @samp{0}.
507 For example, here is the definition used for VAX/VMS:
510 #define INCLUDE_DEFAULTS \
512 @{ "GNU_GXX_INCLUDE:", "G++", 1, 1@}, \
513 @{ "GNU_CC_INCLUDE:", "GCC", 0, 0@}, \
514 @{ "SYS$SYSROOT:[SYSLIB.]", 0, 0, 0@}, \
521 Here is the order of prefixes tried for exec files:
525 Any prefixes specified by the user with @option{-B}.
528 The environment variable @code{GCC_EXEC_PREFIX} or, if @code{GCC_EXEC_PREFIX}
529 is not set and the compiler has not been installed in the configure-time
530 @var{prefix}, the location in which the compiler has actually been installed.
533 The directories specified by the environment variable @code{COMPILER_PATH}.
536 The macro @code{STANDARD_EXEC_PREFIX}, if the compiler has been installed
537 in the configured-time @var{prefix}.
540 The location @file{/usr/libexec/gcc/}, but only if this is a native compiler.
543 The location @file{/usr/lib/gcc/}, but only if this is a native compiler.
546 The macro @code{MD_EXEC_PREFIX}, if defined, but only if this is a native
550 Here is the order of prefixes tried for startfiles:
554 Any prefixes specified by the user with @option{-B}.
557 The environment variable @code{GCC_EXEC_PREFIX} or its automatically determined
558 value based on the installed toolchain location.
561 The directories specified by the environment variable @code{LIBRARY_PATH}
562 (or port-specific name; native only, cross compilers do not use this).
565 The macro @code{STANDARD_EXEC_PREFIX}, but only if the toolchain is installed
566 in the configured @var{prefix} or this is a native compiler.
569 The location @file{/usr/lib/gcc/}, but only if this is a native compiler.
572 The macro @code{MD_EXEC_PREFIX}, if defined, but only if this is a native
576 The macro @code{MD_STARTFILE_PREFIX}, if defined, but only if this is a
577 native compiler, or we have a target system root.
580 The macro @code{MD_STARTFILE_PREFIX_1}, if defined, but only if this is a
581 native compiler, or we have a target system root.
584 The macro @code{STANDARD_STARTFILE_PREFIX}, with any sysroot modifications.
585 If this path is relative it will be prefixed by @code{GCC_EXEC_PREFIX} and
586 the machine suffix or @code{STANDARD_EXEC_PREFIX} and the machine suffix.
589 The macro @code{STANDARD_STARTFILE_PREFIX_1}, but only if this is a native
590 compiler, or we have a target system root. The default for this macro is
594 The macro @code{STANDARD_STARTFILE_PREFIX_2}, but only if this is a native
595 compiler, or we have a target system root. The default for this macro is
599 @node Run-time Target
600 @section Run-time Target Specification
601 @cindex run-time target specification
602 @cindex predefined macros
603 @cindex target specifications
605 @c prevent bad page break with this line
606 Here are run-time target specifications.
608 @defmac TARGET_CPU_CPP_BUILTINS ()
609 This function-like macro expands to a block of code that defines
610 built-in preprocessor macros and assertions for the target CPU, using
611 the functions @code{builtin_define}, @code{builtin_define_std} and
612 @code{builtin_assert}. When the front end
613 calls this macro it provides a trailing semicolon, and since it has
614 finished command line option processing your code can use those
617 @code{builtin_assert} takes a string in the form you pass to the
618 command-line option @option{-A}, such as @code{cpu=mips}, and creates
619 the assertion. @code{builtin_define} takes a string in the form
620 accepted by option @option{-D} and unconditionally defines the macro.
622 @code{builtin_define_std} takes a string representing the name of an
623 object-like macro. If it doesn't lie in the user's namespace,
624 @code{builtin_define_std} defines it unconditionally. Otherwise, it
625 defines a version with two leading underscores, and another version
626 with two leading and trailing underscores, and defines the original
627 only if an ISO standard was not requested on the command line. For
628 example, passing @code{unix} defines @code{__unix}, @code{__unix__}
629 and possibly @code{unix}; passing @code{_mips} defines @code{__mips},
630 @code{__mips__} and possibly @code{_mips}, and passing @code{_ABI64}
631 defines only @code{_ABI64}.
633 You can also test for the C dialect being compiled. The variable
634 @code{c_language} is set to one of @code{clk_c}, @code{clk_cplusplus}
635 or @code{clk_objective_c}. Note that if we are preprocessing
636 assembler, this variable will be @code{clk_c} but the function-like
637 macro @code{preprocessing_asm_p()} will return true, so you might want
638 to check for that first. If you need to check for strict ANSI, the
639 variable @code{flag_iso} can be used. The function-like macro
640 @code{preprocessing_trad_p()} can be used to check for traditional
644 @defmac TARGET_OS_CPP_BUILTINS ()
645 Similarly to @code{TARGET_CPU_CPP_BUILTINS} but this macro is optional
646 and is used for the target operating system instead.
649 @defmac TARGET_OBJFMT_CPP_BUILTINS ()
650 Similarly to @code{TARGET_CPU_CPP_BUILTINS} but this macro is optional
651 and is used for the target object format. @file{elfos.h} uses this
652 macro to define @code{__ELF__}, so you probably do not need to define
656 @deftypevar {extern int} target_flags
657 This variable is declared in @file{options.h}, which is included before
658 any target-specific headers.
661 @hook TARGET_DEFAULT_TARGET_FLAGS
662 This variable specifies the initial value of @code{target_flags}.
663 Its default setting is 0.
666 @cindex optional hardware or system features
667 @cindex features, optional, in system conventions
669 @hook TARGET_HANDLE_OPTION
670 This hook is called whenever the user specifies one of the
671 target-specific options described by the @file{.opt} definition files
672 (@pxref{Options}). It has the opportunity to do some option-specific
673 processing and should return true if the option is valid. The default
674 definition does nothing but return true.
676 @var{decoded} specifies the option and its arguments. @var{opts} and
677 @var{opts_set} are the @code{gcc_options} structures to be used for
678 storing option state, and @var{loc} is the location at which the
679 option was passed (@code{UNKNOWN_LOCATION} except for options passed
683 @hook TARGET_HANDLE_C_OPTION
684 This target hook is called whenever the user specifies one of the
685 target-specific C language family options described by the @file{.opt}
686 definition files(@pxref{Options}). It has the opportunity to do some
687 option-specific processing and should return true if the option is
688 valid. The arguments are like for @code{TARGET_HANDLE_OPTION}. The
689 default definition does nothing but return false.
691 In general, you should use @code{TARGET_HANDLE_OPTION} to handle
692 options. However, if processing an option requires routines that are
693 only available in the C (and related language) front ends, then you
694 should use @code{TARGET_HANDLE_C_OPTION} instead.
697 @hook TARGET_OBJC_CONSTRUCT_STRING_OBJECT
699 @hook TARGET_STRING_OBJECT_REF_TYPE_P
701 @hook TARGET_CHECK_STRING_OBJECT_FORMAT_ARG
703 @hook TARGET_OVERRIDE_OPTIONS_AFTER_CHANGE
704 This target function is similar to the hook @code{TARGET_OPTION_OVERRIDE}
705 but is called when the optimize level is changed via an attribute or
706 pragma or when it is reset at the end of the code affected by the
707 attribute or pragma. It is not called at the beginning of compilation
708 when @code{TARGET_OPTION_OVERRIDE} is called so if you want to perform these
709 actions then, you should have @code{TARGET_OPTION_OVERRIDE} call
710 @code{TARGET_OVERRIDE_OPTIONS_AFTER_CHANGE}.
713 @defmac C_COMMON_OVERRIDE_OPTIONS
714 This is similar to the @code{TARGET_OPTION_OVERRIDE} hook
715 but is only used in the C
716 language frontends (C, Objective-C, C++, Objective-C++) and so can be
717 used to alter option flag variables which only exist in those
721 @hook TARGET_OPTION_OPTIMIZATION_TABLE
722 Some machines may desire to change what optimizations are performed for
723 various optimization levels. This variable, if defined, describes
724 options to enable at particular sets of optimization levels. These
725 options are processed once
726 just after the optimization level is determined and before the remainder
727 of the command options have been parsed, so may be overridden by other
728 options passed explicitly.
730 This processing is run once at program startup and when the optimization
731 options are changed via @code{#pragma GCC optimize} or by using the
732 @code{optimize} attribute.
735 @hook TARGET_OPTION_INIT_STRUCT
737 @hook TARGET_OPTION_DEFAULT_PARAMS
739 @defmac SWITCHABLE_TARGET
740 Some targets need to switch between substantially different subtargets
741 during compilation. For example, the MIPS target has one subtarget for
742 the traditional MIPS architecture and another for MIPS16. Source code
743 can switch between these two subarchitectures using the @code{mips16}
744 and @code{nomips16} attributes.
746 Such subtargets can differ in things like the set of available
747 registers, the set of available instructions, the costs of various
748 operations, and so on. GCC caches a lot of this type of information
749 in global variables, and recomputing them for each subtarget takes a
750 significant amount of time. The compiler therefore provides a facility
751 for maintaining several versions of the global variables and quickly
752 switching between them; see @file{target-globals.h} for details.
754 Define this macro to 1 if your target needs this facility. The default
758 @node Per-Function Data
759 @section Defining data structures for per-function information.
760 @cindex per-function data
761 @cindex data structures
763 If the target needs to store information on a per-function basis, GCC
764 provides a macro and a couple of variables to allow this. Note, just
765 using statics to store the information is a bad idea, since GCC supports
766 nested functions, so you can be halfway through encoding one function
767 when another one comes along.
769 GCC defines a data structure called @code{struct function} which
770 contains all of the data specific to an individual function. This
771 structure contains a field called @code{machine} whose type is
772 @code{struct machine_function *}, which can be used by targets to point
773 to their own specific data.
775 If a target needs per-function specific data it should define the type
776 @code{struct machine_function} and also the macro @code{INIT_EXPANDERS}.
777 This macro should be used to initialize the function pointer
778 @code{init_machine_status}. This pointer is explained below.
780 One typical use of per-function, target specific data is to create an
781 RTX to hold the register containing the function's return address. This
782 RTX can then be used to implement the @code{__builtin_return_address}
783 function, for level 0.
785 Note---earlier implementations of GCC used a single data area to hold
786 all of the per-function information. Thus when processing of a nested
787 function began the old per-function data had to be pushed onto a
788 stack, and when the processing was finished, it had to be popped off the
789 stack. GCC used to provide function pointers called
790 @code{save_machine_status} and @code{restore_machine_status} to handle
791 the saving and restoring of the target specific information. Since the
792 single data area approach is no longer used, these pointers are no
795 @defmac INIT_EXPANDERS
796 Macro called to initialize any target specific information. This macro
797 is called once per function, before generation of any RTL has begun.
798 The intention of this macro is to allow the initialization of the
799 function pointer @code{init_machine_status}.
802 @deftypevar {void (*)(struct function *)} init_machine_status
803 If this function pointer is non-@code{NULL} it will be called once per
804 function, before function compilation starts, in order to allow the
805 target to perform any target specific initialization of the
806 @code{struct function} structure. It is intended that this would be
807 used to initialize the @code{machine} of that structure.
809 @code{struct machine_function} structures are expected to be freed by GC@.
810 Generally, any memory that they reference must be allocated by using
811 GC allocation, including the structure itself.
815 @section Storage Layout
816 @cindex storage layout
818 Note that the definitions of the macros in this table which are sizes or
819 alignments measured in bits do not need to be constant. They can be C
820 expressions that refer to static variables, such as the @code{target_flags}.
821 @xref{Run-time Target}.
823 @defmac BITS_BIG_ENDIAN
824 Define this macro to have the value 1 if the most significant bit in a
825 byte has the lowest number; otherwise define it to have the value zero.
826 This means that bit-field instructions count from the most significant
827 bit. If the machine has no bit-field instructions, then this must still
828 be defined, but it doesn't matter which value it is defined to. This
829 macro need not be a constant.
831 This macro does not affect the way structure fields are packed into
832 bytes or words; that is controlled by @code{BYTES_BIG_ENDIAN}.
835 @defmac BYTES_BIG_ENDIAN
836 Define this macro to have the value 1 if the most significant byte in a
837 word has the lowest number. This macro need not be a constant.
840 @defmac WORDS_BIG_ENDIAN
841 Define this macro to have the value 1 if, in a multiword object, the
842 most significant word has the lowest number. This applies to both
843 memory locations and registers; see @code{REG_WORDS_BIG_ENDIAN} if the
844 order of words in memory is not the same as the order in registers. This
845 macro need not be a constant.
848 @defmac REG_WORDS_BIG_ENDIAN
849 On some machines, the order of words in a multiword object differs between
850 registers in memory. In such a situation, define this macro to describe
851 the order of words in a register. The macro @code{WORDS_BIG_ENDIAN} controls
852 the order of words in memory.
855 @defmac FLOAT_WORDS_BIG_ENDIAN
856 Define this macro to have the value 1 if @code{DFmode}, @code{XFmode} or
857 @code{TFmode} floating point numbers are stored in memory with the word
858 containing the sign bit at the lowest address; otherwise define it to
859 have the value 0. This macro need not be a constant.
861 You need not define this macro if the ordering is the same as for
865 @defmac BITS_PER_UNIT
866 Define this macro to be the number of bits in an addressable storage
867 unit (byte). If you do not define this macro the default is 8.
870 @defmac BITS_PER_WORD
871 Number of bits in a word. If you do not define this macro, the default
872 is @code{BITS_PER_UNIT * UNITS_PER_WORD}.
875 @defmac MAX_BITS_PER_WORD
876 Maximum number of bits in a word. If this is undefined, the default is
877 @code{BITS_PER_WORD}. Otherwise, it is the constant value that is the
878 largest value that @code{BITS_PER_WORD} can have at run-time.
881 @defmac UNITS_PER_WORD
882 Number of storage units in a word; normally the size of a general-purpose
883 register, a power of two from 1 or 8.
886 @defmac MIN_UNITS_PER_WORD
887 Minimum number of units in a word. If this is undefined, the default is
888 @code{UNITS_PER_WORD}. Otherwise, it is the constant value that is the
889 smallest value that @code{UNITS_PER_WORD} can have at run-time.
893 Width of a pointer, in bits. You must specify a value no wider than the
894 width of @code{Pmode}. If it is not equal to the width of @code{Pmode},
895 you must define @code{POINTERS_EXTEND_UNSIGNED}. If you do not specify
896 a value the default is @code{BITS_PER_WORD}.
899 @defmac POINTERS_EXTEND_UNSIGNED
900 A C expression that determines how pointers should be extended from
901 @code{ptr_mode} to either @code{Pmode} or @code{word_mode}. It is
902 greater than zero if pointers should be zero-extended, zero if they
903 should be sign-extended, and negative if some other sort of conversion
904 is needed. In the last case, the extension is done by the target's
905 @code{ptr_extend} instruction.
907 You need not define this macro if the @code{ptr_mode}, @code{Pmode}
908 and @code{word_mode} are all the same width.
911 @defmac PROMOTE_MODE (@var{m}, @var{unsignedp}, @var{type})
912 A macro to update @var{m} and @var{unsignedp} when an object whose type
913 is @var{type} and which has the specified mode and signedness is to be
914 stored in a register. This macro is only called when @var{type} is a
917 On most RISC machines, which only have operations that operate on a full
918 register, define this macro to set @var{m} to @code{word_mode} if
919 @var{m} is an integer mode narrower than @code{BITS_PER_WORD}. In most
920 cases, only integer modes should be widened because wider-precision
921 floating-point operations are usually more expensive than their narrower
924 For most machines, the macro definition does not change @var{unsignedp}.
925 However, some machines, have instructions that preferentially handle
926 either signed or unsigned quantities of certain modes. For example, on
927 the DEC Alpha, 32-bit loads from memory and 32-bit add instructions
928 sign-extend the result to 64 bits. On such machines, set
929 @var{unsignedp} according to which kind of extension is more efficient.
931 Do not define this macro if it would never modify @var{m}.
934 @hook TARGET_PROMOTE_FUNCTION_MODE
935 Like @code{PROMOTE_MODE}, but it is applied to outgoing function arguments or
936 function return values. The target hook should return the new mode
937 and possibly change @code{*@var{punsignedp}} if the promotion should
938 change signedness. This function is called only for scalar @emph{or
941 @var{for_return} allows to distinguish the promotion of arguments and
942 return values. If it is @code{1}, a return value is being promoted and
943 @code{TARGET_FUNCTION_VALUE} must perform the same promotions done here.
944 If it is @code{2}, the returned mode should be that of the register in
945 which an incoming parameter is copied, or the outgoing result is computed;
946 then the hook should return the same mode as @code{promote_mode}, though
947 the signedness may be different.
949 @var{type} can be NULL when promoting function arguments of libcalls.
951 The default is to not promote arguments and return values. You can
952 also define the hook to @code{default_promote_function_mode_always_promote}
953 if you would like to apply the same rules given by @code{PROMOTE_MODE}.
956 @defmac PARM_BOUNDARY
957 Normal alignment required for function parameters on the stack, in
958 bits. All stack parameters receive at least this much alignment
959 regardless of data type. On most machines, this is the same as the
963 @defmac STACK_BOUNDARY
964 Define this macro to the minimum alignment enforced by hardware for the
965 stack pointer on this machine. The definition is a C expression for the
966 desired alignment (measured in bits). This value is used as a default
967 if @code{PREFERRED_STACK_BOUNDARY} is not defined. On most machines,
968 this should be the same as @code{PARM_BOUNDARY}.
971 @defmac PREFERRED_STACK_BOUNDARY
972 Define this macro if you wish to preserve a certain alignment for the
973 stack pointer, greater than what the hardware enforces. The definition
974 is a C expression for the desired alignment (measured in bits). This
975 macro must evaluate to a value equal to or larger than
976 @code{STACK_BOUNDARY}.
979 @defmac INCOMING_STACK_BOUNDARY
980 Define this macro if the incoming stack boundary may be different
981 from @code{PREFERRED_STACK_BOUNDARY}. This macro must evaluate
982 to a value equal to or larger than @code{STACK_BOUNDARY}.
985 @defmac FUNCTION_BOUNDARY
986 Alignment required for a function entry point, in bits.
989 @defmac BIGGEST_ALIGNMENT
990 Biggest alignment that any data type can require on this machine, in
991 bits. Note that this is not the biggest alignment that is supported,
992 just the biggest alignment that, when violated, may cause a fault.
995 @defmac MALLOC_ABI_ALIGNMENT
996 Alignment, in bits, a C conformant malloc implementation has to
997 provide. If not defined, the default value is @code{BITS_PER_WORD}.
1000 @defmac ATTRIBUTE_ALIGNED_VALUE
1001 Alignment used by the @code{__attribute__ ((aligned))} construct. If
1002 not defined, the default value is @code{BIGGEST_ALIGNMENT}.
1005 @defmac MINIMUM_ATOMIC_ALIGNMENT
1006 If defined, the smallest alignment, in bits, that can be given to an
1007 object that can be referenced in one operation, without disturbing any
1008 nearby object. Normally, this is @code{BITS_PER_UNIT}, but may be larger
1009 on machines that don't have byte or half-word store operations.
1012 @defmac BIGGEST_FIELD_ALIGNMENT
1013 Biggest alignment that any structure or union field can require on this
1014 machine, in bits. If defined, this overrides @code{BIGGEST_ALIGNMENT} for
1015 structure and union fields only, unless the field alignment has been set
1016 by the @code{__attribute__ ((aligned (@var{n})))} construct.
1019 @defmac ADJUST_FIELD_ALIGN (@var{field}, @var{computed})
1020 An expression for the alignment of a structure field @var{field} if the
1021 alignment computed in the usual way (including applying of
1022 @code{BIGGEST_ALIGNMENT} and @code{BIGGEST_FIELD_ALIGNMENT} to the
1023 alignment) is @var{computed}. It overrides alignment only if the
1024 field alignment has not been set by the
1025 @code{__attribute__ ((aligned (@var{n})))} construct.
1028 @defmac MAX_STACK_ALIGNMENT
1029 Biggest stack alignment guaranteed by the backend. Use this macro
1030 to specify the maximum alignment of a variable on stack.
1032 If not defined, the default value is @code{STACK_BOUNDARY}.
1034 @c FIXME: The default should be @code{PREFERRED_STACK_BOUNDARY}.
1035 @c But the fix for PR 32893 indicates that we can only guarantee
1036 @c maximum stack alignment on stack up to @code{STACK_BOUNDARY}, not
1037 @c @code{PREFERRED_STACK_BOUNDARY}, if stack alignment isn't supported.
1040 @defmac MAX_OFILE_ALIGNMENT
1041 Biggest alignment supported by the object file format of this machine.
1042 Use this macro to limit the alignment which can be specified using the
1043 @code{__attribute__ ((aligned (@var{n})))} construct. If not defined,
1044 the default value is @code{BIGGEST_ALIGNMENT}.
1046 On systems that use ELF, the default (in @file{config/elfos.h}) is
1047 the largest supported 32-bit ELF section alignment representable on
1048 a 32-bit host e.g. @samp{(((unsigned HOST_WIDEST_INT) 1 << 28) * 8)}.
1049 On 32-bit ELF the largest supported section alignment in bits is
1050 @samp{(0x80000000 * 8)}, but this is not representable on 32-bit hosts.
1053 @defmac DATA_ALIGNMENT (@var{type}, @var{basic-align})
1054 If defined, a C expression to compute the alignment for a variable in
1055 the static store. @var{type} is the data type, and @var{basic-align} is
1056 the alignment that the object would ordinarily have. The value of this
1057 macro is used instead of that alignment to align the object.
1059 If this macro is not defined, then @var{basic-align} is used.
1062 One use of this macro is to increase alignment of medium-size data to
1063 make it all fit in fewer cache lines. Another is to cause character
1064 arrays to be word-aligned so that @code{strcpy} calls that copy
1065 constants to character arrays can be done inline.
1068 @defmac CONSTANT_ALIGNMENT (@var{constant}, @var{basic-align})
1069 If defined, a C expression to compute the alignment given to a constant
1070 that is being placed in memory. @var{constant} is the constant and
1071 @var{basic-align} is the alignment that the object would ordinarily
1072 have. The value of this macro is used instead of that alignment to
1075 If this macro is not defined, then @var{basic-align} is used.
1077 The typical use of this macro is to increase alignment for string
1078 constants to be word aligned so that @code{strcpy} calls that copy
1079 constants can be done inline.
1082 @defmac LOCAL_ALIGNMENT (@var{type}, @var{basic-align})
1083 If defined, a C expression to compute the alignment for a variable in
1084 the local store. @var{type} is the data type, and @var{basic-align} is
1085 the alignment that the object would ordinarily have. The value of this
1086 macro is used instead of that alignment to align the object.
1088 If this macro is not defined, then @var{basic-align} is used.
1090 One use of this macro is to increase alignment of medium-size data to
1091 make it all fit in fewer cache lines.
1093 If the value of this macro has a type, it should be an unsigned type.
1096 @defmac STACK_SLOT_ALIGNMENT (@var{type}, @var{mode}, @var{basic-align})
1097 If defined, a C expression to compute the alignment for stack slot.
1098 @var{type} is the data type, @var{mode} is the widest mode available,
1099 and @var{basic-align} is the alignment that the slot would ordinarily
1100 have. The value of this macro is used instead of that alignment to
1103 If this macro is not defined, then @var{basic-align} is used when
1104 @var{type} is @code{NULL}. Otherwise, @code{LOCAL_ALIGNMENT} will
1107 This macro is to set alignment of stack slot to the maximum alignment
1108 of all possible modes which the slot may have.
1110 If the value of this macro has a type, it should be an unsigned type.
1113 @defmac LOCAL_DECL_ALIGNMENT (@var{decl})
1114 If defined, a C expression to compute the alignment for a local
1115 variable @var{decl}.
1117 If this macro is not defined, then
1118 @code{LOCAL_ALIGNMENT (TREE_TYPE (@var{decl}), DECL_ALIGN (@var{decl}))}
1121 One use of this macro is to increase alignment of medium-size data to
1122 make it all fit in fewer cache lines.
1124 If the value of this macro has a type, it should be an unsigned type.
1127 @defmac MINIMUM_ALIGNMENT (@var{exp}, @var{mode}, @var{align})
1128 If defined, a C expression to compute the minimum required alignment
1129 for dynamic stack realignment purposes for @var{exp} (a type or decl),
1130 @var{mode}, assuming normal alignment @var{align}.
1132 If this macro is not defined, then @var{align} will be used.
1135 @defmac EMPTY_FIELD_BOUNDARY
1136 Alignment in bits to be given to a structure bit-field that follows an
1137 empty field such as @code{int : 0;}.
1139 If @code{PCC_BITFIELD_TYPE_MATTERS} is true, it overrides this macro.
1142 @defmac STRUCTURE_SIZE_BOUNDARY
1143 Number of bits which any structure or union's size must be a multiple of.
1144 Each structure or union's size is rounded up to a multiple of this.
1146 If you do not define this macro, the default is the same as
1147 @code{BITS_PER_UNIT}.
1150 @defmac STRICT_ALIGNMENT
1151 Define this macro to be the value 1 if instructions will fail to work
1152 if given data not on the nominal alignment. If instructions will merely
1153 go slower in that case, define this macro as 0.
1156 @defmac PCC_BITFIELD_TYPE_MATTERS
1157 Define this if you wish to imitate the way many other C compilers handle
1158 alignment of bit-fields and the structures that contain them.
1160 The behavior is that the type written for a named bit-field (@code{int},
1161 @code{short}, or other integer type) imposes an alignment for the entire
1162 structure, as if the structure really did contain an ordinary field of
1163 that type. In addition, the bit-field is placed within the structure so
1164 that it would fit within such a field, not crossing a boundary for it.
1166 Thus, on most machines, a named bit-field whose type is written as
1167 @code{int} would not cross a four-byte boundary, and would force
1168 four-byte alignment for the whole structure. (The alignment used may
1169 not be four bytes; it is controlled by the other alignment parameters.)
1171 An unnamed bit-field will not affect the alignment of the containing
1174 If the macro is defined, its definition should be a C expression;
1175 a nonzero value for the expression enables this behavior.
1177 Note that if this macro is not defined, or its value is zero, some
1178 bit-fields may cross more than one alignment boundary. The compiler can
1179 support such references if there are @samp{insv}, @samp{extv}, and
1180 @samp{extzv} insns that can directly reference memory.
1182 The other known way of making bit-fields work is to define
1183 @code{STRUCTURE_SIZE_BOUNDARY} as large as @code{BIGGEST_ALIGNMENT}.
1184 Then every structure can be accessed with fullwords.
1186 Unless the machine has bit-field instructions or you define
1187 @code{STRUCTURE_SIZE_BOUNDARY} that way, you must define
1188 @code{PCC_BITFIELD_TYPE_MATTERS} to have a nonzero value.
1190 If your aim is to make GCC use the same conventions for laying out
1191 bit-fields as are used by another compiler, here is how to investigate
1192 what the other compiler does. Compile and run this program:
1211 printf ("Size of foo1 is %d\n",
1212 sizeof (struct foo1));
1213 printf ("Size of foo2 is %d\n",
1214 sizeof (struct foo2));
1219 If this prints 2 and 5, then the compiler's behavior is what you would
1220 get from @code{PCC_BITFIELD_TYPE_MATTERS}.
1223 @defmac BITFIELD_NBYTES_LIMITED
1224 Like @code{PCC_BITFIELD_TYPE_MATTERS} except that its effect is limited
1225 to aligning a bit-field within the structure.
1228 @hook TARGET_ALIGN_ANON_BITFIELD
1229 When @code{PCC_BITFIELD_TYPE_MATTERS} is true this hook will determine
1230 whether unnamed bitfields affect the alignment of the containing
1231 structure. The hook should return true if the structure should inherit
1232 the alignment requirements of an unnamed bitfield's type.
1235 @hook TARGET_NARROW_VOLATILE_BITFIELD
1236 This target hook should return @code{true} if accesses to volatile bitfields
1237 should use the narrowest mode possible. It should return @code{false} if
1238 these accesses should use the bitfield container type.
1240 The default is @code{!TARGET_STRICT_ALIGN}.
1243 @defmac MEMBER_TYPE_FORCES_BLK (@var{field}, @var{mode})
1244 Return 1 if a structure or array containing @var{field} should be accessed using
1247 If @var{field} is the only field in the structure, @var{mode} is its
1248 mode, otherwise @var{mode} is VOIDmode. @var{mode} is provided in the
1249 case where structures of one field would require the structure's mode to
1250 retain the field's mode.
1252 Normally, this is not needed.
1255 @defmac ROUND_TYPE_ALIGN (@var{type}, @var{computed}, @var{specified})
1256 Define this macro as an expression for the alignment of a type (given
1257 by @var{type} as a tree node) if the alignment computed in the usual
1258 way is @var{computed} and the alignment explicitly specified was
1261 The default is to use @var{specified} if it is larger; otherwise, use
1262 the smaller of @var{computed} and @code{BIGGEST_ALIGNMENT}
1265 @defmac MAX_FIXED_MODE_SIZE
1266 An integer expression for the size in bits of the largest integer
1267 machine mode that should actually be used. All integer machine modes of
1268 this size or smaller can be used for structures and unions with the
1269 appropriate sizes. If this macro is undefined, @code{GET_MODE_BITSIZE
1270 (DImode)} is assumed.
1273 @defmac STACK_SAVEAREA_MODE (@var{save_level})
1274 If defined, an expression of type @code{enum machine_mode} that
1275 specifies the mode of the save area operand of a
1276 @code{save_stack_@var{level}} named pattern (@pxref{Standard Names}).
1277 @var{save_level} is one of @code{SAVE_BLOCK}, @code{SAVE_FUNCTION}, or
1278 @code{SAVE_NONLOCAL} and selects which of the three named patterns is
1279 having its mode specified.
1281 You need not define this macro if it always returns @code{Pmode}. You
1282 would most commonly define this macro if the
1283 @code{save_stack_@var{level}} patterns need to support both a 32- and a
1287 @defmac STACK_SIZE_MODE
1288 If defined, an expression of type @code{enum machine_mode} that
1289 specifies the mode of the size increment operand of an
1290 @code{allocate_stack} named pattern (@pxref{Standard Names}).
1292 You need not define this macro if it always returns @code{word_mode}.
1293 You would most commonly define this macro if the @code{allocate_stack}
1294 pattern needs to support both a 32- and a 64-bit mode.
1297 @hook TARGET_LIBGCC_CMP_RETURN_MODE
1298 This target hook should return the mode to be used for the return value
1299 of compare instructions expanded to libgcc calls. If not defined
1300 @code{word_mode} is returned which is the right choice for a majority of
1304 @hook TARGET_LIBGCC_SHIFT_COUNT_MODE
1305 This target hook should return the mode to be used for the shift count operand
1306 of shift instructions expanded to libgcc calls. If not defined
1307 @code{word_mode} is returned which is the right choice for a majority of
1311 @hook TARGET_UNWIND_WORD_MODE
1312 Return machine mode to be used for @code{_Unwind_Word} type.
1313 The default is to use @code{word_mode}.
1316 @defmac ROUND_TOWARDS_ZERO
1317 If defined, this macro should be true if the prevailing rounding
1318 mode is towards zero.
1320 Defining this macro only affects the way @file{libgcc.a} emulates
1321 floating-point arithmetic.
1323 Not defining this macro is equivalent to returning zero.
1326 @defmac LARGEST_EXPONENT_IS_NORMAL (@var{size})
1327 This macro should return true if floats with @var{size}
1328 bits do not have a NaN or infinity representation, but use the largest
1329 exponent for normal numbers instead.
1331 Defining this macro only affects the way @file{libgcc.a} emulates
1332 floating-point arithmetic.
1334 The default definition of this macro returns false for all sizes.
1337 @hook TARGET_MS_BITFIELD_LAYOUT_P
1338 This target hook returns @code{true} if bit-fields in the given
1339 @var{record_type} are to be laid out following the rules of Microsoft
1340 Visual C/C++, namely: (i) a bit-field won't share the same storage
1341 unit with the previous bit-field if their underlying types have
1342 different sizes, and the bit-field will be aligned to the highest
1343 alignment of the underlying types of itself and of the previous
1344 bit-field; (ii) a zero-sized bit-field will affect the alignment of
1345 the whole enclosing structure, even if it is unnamed; except that
1346 (iii) a zero-sized bit-field will be disregarded unless it follows
1347 another bit-field of nonzero size. If this hook returns @code{true},
1348 other macros that control bit-field layout are ignored.
1350 When a bit-field is inserted into a packed record, the whole size
1351 of the underlying type is used by one or more same-size adjacent
1352 bit-fields (that is, if its long:3, 32 bits is used in the record,
1353 and any additional adjacent long bit-fields are packed into the same
1354 chunk of 32 bits. However, if the size changes, a new field of that
1355 size is allocated). In an unpacked record, this is the same as using
1356 alignment, but not equivalent when packing.
1358 If both MS bit-fields and @samp{__attribute__((packed))} are used,
1359 the latter will take precedence. If @samp{__attribute__((packed))} is
1360 used on a single field when MS bit-fields are in use, it will take
1361 precedence for that field, but the alignment of the rest of the structure
1362 may affect its placement.
1365 @hook TARGET_DECIMAL_FLOAT_SUPPORTED_P
1366 Returns true if the target supports decimal floating point.
1369 @hook TARGET_FIXED_POINT_SUPPORTED_P
1370 Returns true if the target supports fixed-point arithmetic.
1373 @hook TARGET_EXPAND_TO_RTL_HOOK
1374 This hook is called just before expansion into rtl, allowing the target
1375 to perform additional initializations or analysis before the expansion.
1376 For example, the rs6000 port uses it to allocate a scratch stack slot
1377 for use in copying SDmode values between memory and floating point
1378 registers whenever the function being expanded has any SDmode
1382 @hook TARGET_INSTANTIATE_DECLS
1383 This hook allows the backend to perform additional instantiations on rtl
1384 that are not actually in any insns yet, but will be later.
1387 @hook TARGET_MANGLE_TYPE
1388 If your target defines any fundamental types, or any types your target
1389 uses should be mangled differently from the default, define this hook
1390 to return the appropriate encoding for these types as part of a C++
1391 mangled name. The @var{type} argument is the tree structure representing
1392 the type to be mangled. The hook may be applied to trees which are
1393 not target-specific fundamental types; it should return @code{NULL}
1394 for all such types, as well as arguments it does not recognize. If the
1395 return value is not @code{NULL}, it must point to a statically-allocated
1398 Target-specific fundamental types might be new fundamental types or
1399 qualified versions of ordinary fundamental types. Encode new
1400 fundamental types as @samp{@w{u @var{n} @var{name}}}, where @var{name}
1401 is the name used for the type in source code, and @var{n} is the
1402 length of @var{name} in decimal. Encode qualified versions of
1403 ordinary types as @samp{@w{U @var{n} @var{name} @var{code}}}, where
1404 @var{name} is the name used for the type qualifier in source code,
1405 @var{n} is the length of @var{name} as above, and @var{code} is the
1406 code used to represent the unqualified version of this type. (See
1407 @code{write_builtin_type} in @file{cp/mangle.c} for the list of
1408 codes.) In both cases the spaces are for clarity; do not include any
1409 spaces in your string.
1411 This hook is applied to types prior to typedef resolution. If the mangled
1412 name for a particular type depends only on that type's main variant, you
1413 can perform typedef resolution yourself using @code{TYPE_MAIN_VARIANT}
1416 The default version of this hook always returns @code{NULL}, which is
1417 appropriate for a target that does not define any new fundamental
1422 @section Layout of Source Language Data Types
1424 These macros define the sizes and other characteristics of the standard
1425 basic data types used in programs being compiled. Unlike the macros in
1426 the previous section, these apply to specific features of C and related
1427 languages, rather than to fundamental aspects of storage layout.
1429 @defmac INT_TYPE_SIZE
1430 A C expression for the size in bits of the type @code{int} on the
1431 target machine. If you don't define this, the default is one word.
1434 @defmac SHORT_TYPE_SIZE
1435 A C expression for the size in bits of the type @code{short} on the
1436 target machine. If you don't define this, the default is half a word.
1437 (If this would be less than one storage unit, it is rounded up to one
1441 @defmac LONG_TYPE_SIZE
1442 A C expression for the size in bits of the type @code{long} on the
1443 target machine. If you don't define this, the default is one word.
1446 @defmac ADA_LONG_TYPE_SIZE
1447 On some machines, the size used for the Ada equivalent of the type
1448 @code{long} by a native Ada compiler differs from that used by C@. In
1449 that situation, define this macro to be a C expression to be used for
1450 the size of that type. If you don't define this, the default is the
1451 value of @code{LONG_TYPE_SIZE}.
1454 @defmac LONG_LONG_TYPE_SIZE
1455 A C expression for the size in bits of the type @code{long long} on the
1456 target machine. If you don't define this, the default is two
1457 words. If you want to support GNU Ada on your machine, the value of this
1458 macro must be at least 64.
1461 @defmac CHAR_TYPE_SIZE
1462 A C expression for the size in bits of the type @code{char} on the
1463 target machine. If you don't define this, the default is
1464 @code{BITS_PER_UNIT}.
1467 @defmac BOOL_TYPE_SIZE
1468 A C expression for the size in bits of the C++ type @code{bool} and
1469 C99 type @code{_Bool} on the target machine. If you don't define
1470 this, and you probably shouldn't, the default is @code{CHAR_TYPE_SIZE}.
1473 @defmac FLOAT_TYPE_SIZE
1474 A C expression for the size in bits of the type @code{float} on the
1475 target machine. If you don't define this, the default is one word.
1478 @defmac DOUBLE_TYPE_SIZE
1479 A C expression for the size in bits of the type @code{double} on the
1480 target machine. If you don't define this, the default is two
1484 @defmac LONG_DOUBLE_TYPE_SIZE
1485 A C expression for the size in bits of the type @code{long double} on
1486 the target machine. If you don't define this, the default is two
1490 @defmac SHORT_FRACT_TYPE_SIZE
1491 A C expression for the size in bits of the type @code{short _Fract} on
1492 the target machine. If you don't define this, the default is
1493 @code{BITS_PER_UNIT}.
1496 @defmac FRACT_TYPE_SIZE
1497 A C expression for the size in bits of the type @code{_Fract} on
1498 the target machine. If you don't define this, the default is
1499 @code{BITS_PER_UNIT * 2}.
1502 @defmac LONG_FRACT_TYPE_SIZE
1503 A C expression for the size in bits of the type @code{long _Fract} on
1504 the target machine. If you don't define this, the default is
1505 @code{BITS_PER_UNIT * 4}.
1508 @defmac LONG_LONG_FRACT_TYPE_SIZE
1509 A C expression for the size in bits of the type @code{long long _Fract} on
1510 the target machine. If you don't define this, the default is
1511 @code{BITS_PER_UNIT * 8}.
1514 @defmac SHORT_ACCUM_TYPE_SIZE
1515 A C expression for the size in bits of the type @code{short _Accum} on
1516 the target machine. If you don't define this, the default is
1517 @code{BITS_PER_UNIT * 2}.
1520 @defmac ACCUM_TYPE_SIZE
1521 A C expression for the size in bits of the type @code{_Accum} on
1522 the target machine. If you don't define this, the default is
1523 @code{BITS_PER_UNIT * 4}.
1526 @defmac LONG_ACCUM_TYPE_SIZE
1527 A C expression for the size in bits of the type @code{long _Accum} on
1528 the target machine. If you don't define this, the default is
1529 @code{BITS_PER_UNIT * 8}.
1532 @defmac LONG_LONG_ACCUM_TYPE_SIZE
1533 A C expression for the size in bits of the type @code{long long _Accum} on
1534 the target machine. If you don't define this, the default is
1535 @code{BITS_PER_UNIT * 16}.
1538 @defmac LIBGCC2_LONG_DOUBLE_TYPE_SIZE
1539 Define this macro if @code{LONG_DOUBLE_TYPE_SIZE} is not constant or
1540 if you want routines in @file{libgcc2.a} for a size other than
1541 @code{LONG_DOUBLE_TYPE_SIZE}. If you don't define this, the
1542 default is @code{LONG_DOUBLE_TYPE_SIZE}.
1545 @defmac LIBGCC2_HAS_DF_MODE
1546 Define this macro if neither @code{DOUBLE_TYPE_SIZE} nor
1547 @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is
1548 @code{DFmode} but you want @code{DFmode} routines in @file{libgcc2.a}
1549 anyway. If you don't define this and either @code{DOUBLE_TYPE_SIZE}
1550 or @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is 64 then the default is 1,
1554 @defmac LIBGCC2_HAS_XF_MODE
1555 Define this macro if @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is not
1556 @code{XFmode} but you want @code{XFmode} routines in @file{libgcc2.a}
1557 anyway. If you don't define this and @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE}
1558 is 80 then the default is 1, otherwise it is 0.
1561 @defmac LIBGCC2_HAS_TF_MODE
1562 Define this macro if @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is not
1563 @code{TFmode} but you want @code{TFmode} routines in @file{libgcc2.a}
1564 anyway. If you don't define this and @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE}
1565 is 128 then the default is 1, otherwise it is 0.
1568 @defmac LIBGCC2_GNU_PREFIX
1569 This macro corresponds to the @code{TARGET_LIBFUNC_GNU_PREFIX} target
1570 hook and should be defined if that hook is overriden to be true. It
1571 causes function names in libgcc to be changed to use a @code{__gnu_}
1572 prefix for their name rather than the default @code{__}. A port which
1573 uses this macro should also arrange to use @file{t-gnu-prefix} in
1574 the libgcc @file{config.host}.
1581 Define these macros to be the size in bits of the mantissa of
1582 @code{SFmode}, @code{DFmode}, @code{XFmode} and @code{TFmode} values,
1583 if the defaults in @file{libgcc2.h} are inappropriate. By default,
1584 @code{FLT_MANT_DIG} is used for @code{SF_SIZE}, @code{LDBL_MANT_DIG}
1585 for @code{XF_SIZE} and @code{TF_SIZE}, and @code{DBL_MANT_DIG} or
1586 @code{LDBL_MANT_DIG} for @code{DF_SIZE} according to whether
1587 @code{DOUBLE_TYPE_SIZE} or
1588 @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is 64.
1591 @defmac TARGET_FLT_EVAL_METHOD
1592 A C expression for the value for @code{FLT_EVAL_METHOD} in @file{float.h},
1593 assuming, if applicable, that the floating-point control word is in its
1594 default state. If you do not define this macro the value of
1595 @code{FLT_EVAL_METHOD} will be zero.
1598 @defmac WIDEST_HARDWARE_FP_SIZE
1599 A C expression for the size in bits of the widest floating-point format
1600 supported by the hardware. If you define this macro, you must specify a
1601 value less than or equal to the value of @code{LONG_DOUBLE_TYPE_SIZE}.
1602 If you do not define this macro, the value of @code{LONG_DOUBLE_TYPE_SIZE}
1606 @defmac DEFAULT_SIGNED_CHAR
1607 An expression whose value is 1 or 0, according to whether the type
1608 @code{char} should be signed or unsigned by default. The user can
1609 always override this default with the options @option{-fsigned-char}
1610 and @option{-funsigned-char}.
1613 @hook TARGET_DEFAULT_SHORT_ENUMS
1614 This target hook should return true if the compiler should give an
1615 @code{enum} type only as many bytes as it takes to represent the range
1616 of possible values of that type. It should return false if all
1617 @code{enum} types should be allocated like @code{int}.
1619 The default is to return false.
1623 A C expression for a string describing the name of the data type to use
1624 for size values. The typedef name @code{size_t} is defined using the
1625 contents of the string.
1627 The string can contain more than one keyword. If so, separate them with
1628 spaces, and write first any length keyword, then @code{unsigned} if
1629 appropriate, and finally @code{int}. The string must exactly match one
1630 of the data type names defined in the function
1631 @code{init_decl_processing} in the file @file{c-decl.c}. You may not
1632 omit @code{int} or change the order---that would cause the compiler to
1635 If you don't define this macro, the default is @code{"long unsigned
1639 @defmac PTRDIFF_TYPE
1640 A C expression for a string describing the name of the data type to use
1641 for the result of subtracting two pointers. The typedef name
1642 @code{ptrdiff_t} is defined using the contents of the string. See
1643 @code{SIZE_TYPE} above for more information.
1645 If you don't define this macro, the default is @code{"long int"}.
1649 A C expression for a string describing the name of the data type to use
1650 for wide characters. The typedef name @code{wchar_t} is defined using
1651 the contents of the string. See @code{SIZE_TYPE} above for more
1654 If you don't define this macro, the default is @code{"int"}.
1657 @defmac WCHAR_TYPE_SIZE
1658 A C expression for the size in bits of the data type for wide
1659 characters. This is used in @code{cpp}, which cannot make use of
1664 A C expression for a string describing the name of the data type to
1665 use for wide characters passed to @code{printf} and returned from
1666 @code{getwc}. The typedef name @code{wint_t} is defined using the
1667 contents of the string. See @code{SIZE_TYPE} above for more
1670 If you don't define this macro, the default is @code{"unsigned int"}.
1674 A C expression for a string describing the name of the data type that
1675 can represent any value of any standard or extended signed integer type.
1676 The typedef name @code{intmax_t} is defined using the contents of the
1677 string. See @code{SIZE_TYPE} above for more information.
1679 If you don't define this macro, the default is the first of
1680 @code{"int"}, @code{"long int"}, or @code{"long long int"} that has as
1681 much precision as @code{long long int}.
1684 @defmac UINTMAX_TYPE
1685 A C expression for a string describing the name of the data type that
1686 can represent any value of any standard or extended unsigned integer
1687 type. The typedef name @code{uintmax_t} is defined using the contents
1688 of the string. See @code{SIZE_TYPE} above for more information.
1690 If you don't define this macro, the default is the first of
1691 @code{"unsigned int"}, @code{"long unsigned int"}, or @code{"long long
1692 unsigned int"} that has as much precision as @code{long long unsigned
1696 @defmac SIG_ATOMIC_TYPE
1702 @defmacx UINT16_TYPE
1703 @defmacx UINT32_TYPE
1704 @defmacx UINT64_TYPE
1705 @defmacx INT_LEAST8_TYPE
1706 @defmacx INT_LEAST16_TYPE
1707 @defmacx INT_LEAST32_TYPE
1708 @defmacx INT_LEAST64_TYPE
1709 @defmacx UINT_LEAST8_TYPE
1710 @defmacx UINT_LEAST16_TYPE
1711 @defmacx UINT_LEAST32_TYPE
1712 @defmacx UINT_LEAST64_TYPE
1713 @defmacx INT_FAST8_TYPE
1714 @defmacx INT_FAST16_TYPE
1715 @defmacx INT_FAST32_TYPE
1716 @defmacx INT_FAST64_TYPE
1717 @defmacx UINT_FAST8_TYPE
1718 @defmacx UINT_FAST16_TYPE
1719 @defmacx UINT_FAST32_TYPE
1720 @defmacx UINT_FAST64_TYPE
1721 @defmacx INTPTR_TYPE
1722 @defmacx UINTPTR_TYPE
1723 C expressions for the standard types @code{sig_atomic_t},
1724 @code{int8_t}, @code{int16_t}, @code{int32_t}, @code{int64_t},
1725 @code{uint8_t}, @code{uint16_t}, @code{uint32_t}, @code{uint64_t},
1726 @code{int_least8_t}, @code{int_least16_t}, @code{int_least32_t},
1727 @code{int_least64_t}, @code{uint_least8_t}, @code{uint_least16_t},
1728 @code{uint_least32_t}, @code{uint_least64_t}, @code{int_fast8_t},
1729 @code{int_fast16_t}, @code{int_fast32_t}, @code{int_fast64_t},
1730 @code{uint_fast8_t}, @code{uint_fast16_t}, @code{uint_fast32_t},
1731 @code{uint_fast64_t}, @code{intptr_t}, and @code{uintptr_t}. See
1732 @code{SIZE_TYPE} above for more information.
1734 If any of these macros evaluates to a null pointer, the corresponding
1735 type is not supported; if GCC is configured to provide
1736 @code{<stdint.h>} in such a case, the header provided may not conform
1737 to C99, depending on the type in question. The defaults for all of
1738 these macros are null pointers.
1741 @defmac TARGET_PTRMEMFUNC_VBIT_LOCATION
1742 The C++ compiler represents a pointer-to-member-function with a struct
1749 ptrdiff_t vtable_index;
1756 The C++ compiler must use one bit to indicate whether the function that
1757 will be called through a pointer-to-member-function is virtual.
1758 Normally, we assume that the low-order bit of a function pointer must
1759 always be zero. Then, by ensuring that the vtable_index is odd, we can
1760 distinguish which variant of the union is in use. But, on some
1761 platforms function pointers can be odd, and so this doesn't work. In
1762 that case, we use the low-order bit of the @code{delta} field, and shift
1763 the remainder of the @code{delta} field to the left.
1765 GCC will automatically make the right selection about where to store
1766 this bit using the @code{FUNCTION_BOUNDARY} setting for your platform.
1767 However, some platforms such as ARM/Thumb have @code{FUNCTION_BOUNDARY}
1768 set such that functions always start at even addresses, but the lowest
1769 bit of pointers to functions indicate whether the function at that
1770 address is in ARM or Thumb mode. If this is the case of your
1771 architecture, you should define this macro to
1772 @code{ptrmemfunc_vbit_in_delta}.
1774 In general, you should not have to define this macro. On architectures
1775 in which function addresses are always even, according to
1776 @code{FUNCTION_BOUNDARY}, GCC will automatically define this macro to
1777 @code{ptrmemfunc_vbit_in_pfn}.
1780 @defmac TARGET_VTABLE_USES_DESCRIPTORS
1781 Normally, the C++ compiler uses function pointers in vtables. This
1782 macro allows the target to change to use ``function descriptors''
1783 instead. Function descriptors are found on targets for whom a
1784 function pointer is actually a small data structure. Normally the
1785 data structure consists of the actual code address plus a data
1786 pointer to which the function's data is relative.
1788 If vtables are used, the value of this macro should be the number
1789 of words that the function descriptor occupies.
1792 @defmac TARGET_VTABLE_ENTRY_ALIGN
1793 By default, the vtable entries are void pointers, the so the alignment
1794 is the same as pointer alignment. The value of this macro specifies
1795 the alignment of the vtable entry in bits. It should be defined only
1796 when special alignment is necessary. */
1799 @defmac TARGET_VTABLE_DATA_ENTRY_DISTANCE
1800 There are a few non-descriptor entries in the vtable at offsets below
1801 zero. If these entries must be padded (say, to preserve the alignment
1802 specified by @code{TARGET_VTABLE_ENTRY_ALIGN}), set this to the number
1803 of words in each data entry.
1807 @section Register Usage
1808 @cindex register usage
1810 This section explains how to describe what registers the target machine
1811 has, and how (in general) they can be used.
1813 The description of which registers a specific instruction can use is
1814 done with register classes; see @ref{Register Classes}. For information
1815 on using registers to access a stack frame, see @ref{Frame Registers}.
1816 For passing values in registers, see @ref{Register Arguments}.
1817 For returning values in registers, see @ref{Scalar Return}.
1820 * Register Basics:: Number and kinds of registers.
1821 * Allocation Order:: Order in which registers are allocated.
1822 * Values in Registers:: What kinds of values each reg can hold.
1823 * Leaf Functions:: Renumbering registers for leaf functions.
1824 * Stack Registers:: Handling a register stack such as 80387.
1827 @node Register Basics
1828 @subsection Basic Characteristics of Registers
1830 @c prevent bad page break with this line
1831 Registers have various characteristics.
1833 @defmac FIRST_PSEUDO_REGISTER
1834 Number of hardware registers known to the compiler. They receive
1835 numbers 0 through @code{FIRST_PSEUDO_REGISTER-1}; thus, the first
1836 pseudo register's number really is assigned the number
1837 @code{FIRST_PSEUDO_REGISTER}.
1840 @defmac FIXED_REGISTERS
1841 @cindex fixed register
1842 An initializer that says which registers are used for fixed purposes
1843 all throughout the compiled code and are therefore not available for
1844 general allocation. These would include the stack pointer, the frame
1845 pointer (except on machines where that can be used as a general
1846 register when no frame pointer is needed), the program counter on
1847 machines where that is considered one of the addressable registers,
1848 and any other numbered register with a standard use.
1850 This information is expressed as a sequence of numbers, separated by
1851 commas and surrounded by braces. The @var{n}th number is 1 if
1852 register @var{n} is fixed, 0 otherwise.
1854 The table initialized from this macro, and the table initialized by
1855 the following one, may be overridden at run time either automatically,
1856 by the actions of the macro @code{CONDITIONAL_REGISTER_USAGE}, or by
1857 the user with the command options @option{-ffixed-@var{reg}},
1858 @option{-fcall-used-@var{reg}} and @option{-fcall-saved-@var{reg}}.
1861 @defmac CALL_USED_REGISTERS
1862 @cindex call-used register
1863 @cindex call-clobbered register
1864 @cindex call-saved register
1865 Like @code{FIXED_REGISTERS} but has 1 for each register that is
1866 clobbered (in general) by function calls as well as for fixed
1867 registers. This macro therefore identifies the registers that are not
1868 available for general allocation of values that must live across
1871 If a register has 0 in @code{CALL_USED_REGISTERS}, the compiler
1872 automatically saves it on function entry and restores it on function
1873 exit, if the register is used within the function.
1876 @defmac CALL_REALLY_USED_REGISTERS
1877 @cindex call-used register
1878 @cindex call-clobbered register
1879 @cindex call-saved register
1880 Like @code{CALL_USED_REGISTERS} except this macro doesn't require
1881 that the entire set of @code{FIXED_REGISTERS} be included.
1882 (@code{CALL_USED_REGISTERS} must be a superset of @code{FIXED_REGISTERS}).
1883 This macro is optional. If not specified, it defaults to the value
1884 of @code{CALL_USED_REGISTERS}.
1887 @defmac HARD_REGNO_CALL_PART_CLOBBERED (@var{regno}, @var{mode})
1888 @cindex call-used register
1889 @cindex call-clobbered register
1890 @cindex call-saved register
1891 A C expression that is nonzero if it is not permissible to store a
1892 value of mode @var{mode} in hard register number @var{regno} across a
1893 call without some part of it being clobbered. For most machines this
1894 macro need not be defined. It is only required for machines that do not
1895 preserve the entire contents of a register across a call.
1899 @findex call_used_regs
1902 @findex reg_class_contents
1903 @hook TARGET_CONDITIONAL_REGISTER_USAGE
1904 This hook may conditionally modify five variables
1905 @code{fixed_regs}, @code{call_used_regs}, @code{global_regs},
1906 @code{reg_names}, and @code{reg_class_contents}, to take into account
1907 any dependence of these register sets on target flags. The first three
1908 of these are of type @code{char []} (interpreted as Boolean vectors).
1909 @code{global_regs} is a @code{const char *[]}, and
1910 @code{reg_class_contents} is a @code{HARD_REG_SET}. Before the macro is
1911 called, @code{fixed_regs}, @code{call_used_regs},
1912 @code{reg_class_contents}, and @code{reg_names} have been initialized
1913 from @code{FIXED_REGISTERS}, @code{CALL_USED_REGISTERS},
1914 @code{REG_CLASS_CONTENTS}, and @code{REGISTER_NAMES}, respectively.
1915 @code{global_regs} has been cleared, and any @option{-ffixed-@var{reg}},
1916 @option{-fcall-used-@var{reg}} and @option{-fcall-saved-@var{reg}}
1917 command options have been applied.
1919 @cindex disabling certain registers
1920 @cindex controlling register usage
1921 If the usage of an entire class of registers depends on the target
1922 flags, you may indicate this to GCC by using this macro to modify
1923 @code{fixed_regs} and @code{call_used_regs} to 1 for each of the
1924 registers in the classes which should not be used by GCC@. Also define
1925 the macro @code{REG_CLASS_FROM_LETTER} / @code{REG_CLASS_FROM_CONSTRAINT}
1926 to return @code{NO_REGS} if it
1927 is called with a letter for a class that shouldn't be used.
1929 (However, if this class is not included in @code{GENERAL_REGS} and all
1930 of the insn patterns whose constraints permit this class are
1931 controlled by target switches, then GCC will automatically avoid using
1932 these registers when the target switches are opposed to them.)
1935 @defmac INCOMING_REGNO (@var{out})
1936 Define this macro if the target machine has register windows. This C
1937 expression returns the register number as seen by the called function
1938 corresponding to the register number @var{out} as seen by the calling
1939 function. Return @var{out} if register number @var{out} is not an
1943 @defmac OUTGOING_REGNO (@var{in})
1944 Define this macro if the target machine has register windows. This C
1945 expression returns the register number as seen by the calling function
1946 corresponding to the register number @var{in} as seen by the called
1947 function. Return @var{in} if register number @var{in} is not an inbound
1951 @defmac LOCAL_REGNO (@var{regno})
1952 Define this macro if the target machine has register windows. This C
1953 expression returns true if the register is call-saved but is in the
1954 register window. Unlike most call-saved registers, such registers
1955 need not be explicitly restored on function exit or during non-local
1960 If the program counter has a register number, define this as that
1961 register number. Otherwise, do not define it.
1964 @node Allocation Order
1965 @subsection Order of Allocation of Registers
1966 @cindex order of register allocation
1967 @cindex register allocation order
1969 @c prevent bad page break with this line
1970 Registers are allocated in order.
1972 @defmac REG_ALLOC_ORDER
1973 If defined, an initializer for a vector of integers, containing the
1974 numbers of hard registers in the order in which GCC should prefer
1975 to use them (from most preferred to least).
1977 If this macro is not defined, registers are used lowest numbered first
1978 (all else being equal).
1980 One use of this macro is on machines where the highest numbered
1981 registers must always be saved and the save-multiple-registers
1982 instruction supports only sequences of consecutive registers. On such
1983 machines, define @code{REG_ALLOC_ORDER} to be an initializer that lists
1984 the highest numbered allocable register first.
1987 @defmac ADJUST_REG_ALLOC_ORDER
1988 A C statement (sans semicolon) to choose the order in which to allocate
1989 hard registers for pseudo-registers local to a basic block.
1991 Store the desired register order in the array @code{reg_alloc_order}.
1992 Element 0 should be the register to allocate first; element 1, the next
1993 register; and so on.
1995 The macro body should not assume anything about the contents of
1996 @code{reg_alloc_order} before execution of the macro.
1998 On most machines, it is not necessary to define this macro.
2001 @defmac HONOR_REG_ALLOC_ORDER
2002 Normally, IRA tries to estimate the costs for saving a register in the
2003 prologue and restoring it in the epilogue. This discourages it from
2004 using call-saved registers. If a machine wants to ensure that IRA
2005 allocates registers in the order given by REG_ALLOC_ORDER even if some
2006 call-saved registers appear earlier than call-used ones, this macro
2010 @defmac IRA_HARD_REGNO_ADD_COST_MULTIPLIER (@var{regno})
2011 In some case register allocation order is not enough for the
2012 Integrated Register Allocator (@acronym{IRA}) to generate a good code.
2013 If this macro is defined, it should return a floating point value
2014 based on @var{regno}. The cost of using @var{regno} for a pseudo will
2015 be increased by approximately the pseudo's usage frequency times the
2016 value returned by this macro. Not defining this macro is equivalent
2017 to having it always return @code{0.0}.
2019 On most machines, it is not necessary to define this macro.
2022 @node Values in Registers
2023 @subsection How Values Fit in Registers
2025 This section discusses the macros that describe which kinds of values
2026 (specifically, which machine modes) each register can hold, and how many
2027 consecutive registers are needed for a given mode.
2029 @defmac HARD_REGNO_NREGS (@var{regno}, @var{mode})
2030 A C expression for the number of consecutive hard registers, starting
2031 at register number @var{regno}, required to hold a value of mode
2032 @var{mode}. This macro must never return zero, even if a register
2033 cannot hold the requested mode - indicate that with HARD_REGNO_MODE_OK
2034 and/or CANNOT_CHANGE_MODE_CLASS instead.
2036 On a machine where all registers are exactly one word, a suitable
2037 definition of this macro is
2040 #define HARD_REGNO_NREGS(REGNO, MODE) \
2041 ((GET_MODE_SIZE (MODE) + UNITS_PER_WORD - 1) \
2046 @defmac HARD_REGNO_NREGS_HAS_PADDING (@var{regno}, @var{mode})
2047 A C expression that is nonzero if a value of mode @var{mode}, stored
2048 in memory, ends with padding that causes it to take up more space than
2049 in registers starting at register number @var{regno} (as determined by
2050 multiplying GCC's notion of the size of the register when containing
2051 this mode by the number of registers returned by
2052 @code{HARD_REGNO_NREGS}). By default this is zero.
2054 For example, if a floating-point value is stored in three 32-bit
2055 registers but takes up 128 bits in memory, then this would be
2058 This macros only needs to be defined if there are cases where
2059 @code{subreg_get_info}
2060 would otherwise wrongly determine that a @code{subreg} can be
2061 represented by an offset to the register number, when in fact such a
2062 @code{subreg} would contain some of the padding not stored in
2063 registers and so not be representable.
2066 @defmac HARD_REGNO_NREGS_WITH_PADDING (@var{regno}, @var{mode})
2067 For values of @var{regno} and @var{mode} for which
2068 @code{HARD_REGNO_NREGS_HAS_PADDING} returns nonzero, a C expression
2069 returning the greater number of registers required to hold the value
2070 including any padding. In the example above, the value would be four.
2073 @defmac REGMODE_NATURAL_SIZE (@var{mode})
2074 Define this macro if the natural size of registers that hold values
2075 of mode @var{mode} is not the word size. It is a C expression that
2076 should give the natural size in bytes for the specified mode. It is
2077 used by the register allocator to try to optimize its results. This
2078 happens for example on SPARC 64-bit where the natural size of
2079 floating-point registers is still 32-bit.
2082 @defmac HARD_REGNO_MODE_OK (@var{regno}, @var{mode})
2083 A C expression that is nonzero if it is permissible to store a value
2084 of mode @var{mode} in hard register number @var{regno} (or in several
2085 registers starting with that one). For a machine where all registers
2086 are equivalent, a suitable definition is
2089 #define HARD_REGNO_MODE_OK(REGNO, MODE) 1
2092 You need not include code to check for the numbers of fixed registers,
2093 because the allocation mechanism considers them to be always occupied.
2095 @cindex register pairs
2096 On some machines, double-precision values must be kept in even/odd
2097 register pairs. You can implement that by defining this macro to reject
2098 odd register numbers for such modes.
2100 The minimum requirement for a mode to be OK in a register is that the
2101 @samp{mov@var{mode}} instruction pattern support moves between the
2102 register and other hard register in the same class and that moving a
2103 value into the register and back out not alter it.
2105 Since the same instruction used to move @code{word_mode} will work for
2106 all narrower integer modes, it is not necessary on any machine for
2107 @code{HARD_REGNO_MODE_OK} to distinguish between these modes, provided
2108 you define patterns @samp{movhi}, etc., to take advantage of this. This
2109 is useful because of the interaction between @code{HARD_REGNO_MODE_OK}
2110 and @code{MODES_TIEABLE_P}; it is very desirable for all integer modes
2113 Many machines have special registers for floating point arithmetic.
2114 Often people assume that floating point machine modes are allowed only
2115 in floating point registers. This is not true. Any registers that
2116 can hold integers can safely @emph{hold} a floating point machine
2117 mode, whether or not floating arithmetic can be done on it in those
2118 registers. Integer move instructions can be used to move the values.
2120 On some machines, though, the converse is true: fixed-point machine
2121 modes may not go in floating registers. This is true if the floating
2122 registers normalize any value stored in them, because storing a
2123 non-floating value there would garble it. In this case,
2124 @code{HARD_REGNO_MODE_OK} should reject fixed-point machine modes in
2125 floating registers. But if the floating registers do not automatically
2126 normalize, if you can store any bit pattern in one and retrieve it
2127 unchanged without a trap, then any machine mode may go in a floating
2128 register, so you can define this macro to say so.
2130 The primary significance of special floating registers is rather that
2131 they are the registers acceptable in floating point arithmetic
2132 instructions. However, this is of no concern to
2133 @code{HARD_REGNO_MODE_OK}. You handle it by writing the proper
2134 constraints for those instructions.
2136 On some machines, the floating registers are especially slow to access,
2137 so that it is better to store a value in a stack frame than in such a
2138 register if floating point arithmetic is not being done. As long as the
2139 floating registers are not in class @code{GENERAL_REGS}, they will not
2140 be used unless some pattern's constraint asks for one.
2143 @defmac HARD_REGNO_RENAME_OK (@var{from}, @var{to})
2144 A C expression that is nonzero if it is OK to rename a hard register
2145 @var{from} to another hard register @var{to}.
2147 One common use of this macro is to prevent renaming of a register to
2148 another register that is not saved by a prologue in an interrupt
2151 The default is always nonzero.
2154 @defmac MODES_TIEABLE_P (@var{mode1}, @var{mode2})
2155 A C expression that is nonzero if a value of mode
2156 @var{mode1} is accessible in mode @var{mode2} without copying.
2158 If @code{HARD_REGNO_MODE_OK (@var{r}, @var{mode1})} and
2159 @code{HARD_REGNO_MODE_OK (@var{r}, @var{mode2})} are always the same for
2160 any @var{r}, then @code{MODES_TIEABLE_P (@var{mode1}, @var{mode2})}
2161 should be nonzero. If they differ for any @var{r}, you should define
2162 this macro to return zero unless some other mechanism ensures the
2163 accessibility of the value in a narrower mode.
2165 You should define this macro to return nonzero in as many cases as
2166 possible since doing so will allow GCC to perform better register
2170 @hook TARGET_HARD_REGNO_SCRATCH_OK
2171 This target hook should return @code{true} if it is OK to use a hard register
2172 @var{regno} as scratch reg in peephole2.
2174 One common use of this macro is to prevent using of a register that
2175 is not saved by a prologue in an interrupt handler.
2177 The default version of this hook always returns @code{true}.
2180 @defmac AVOID_CCMODE_COPIES
2181 Define this macro if the compiler should avoid copies to/from @code{CCmode}
2182 registers. You should only define this macro if support for copying to/from
2183 @code{CCmode} is incomplete.
2186 @node Leaf Functions
2187 @subsection Handling Leaf Functions
2189 @cindex leaf functions
2190 @cindex functions, leaf
2191 On some machines, a leaf function (i.e., one which makes no calls) can run
2192 more efficiently if it does not make its own register window. Often this
2193 means it is required to receive its arguments in the registers where they
2194 are passed by the caller, instead of the registers where they would
2197 The special treatment for leaf functions generally applies only when
2198 other conditions are met; for example, often they may use only those
2199 registers for its own variables and temporaries. We use the term ``leaf
2200 function'' to mean a function that is suitable for this special
2201 handling, so that functions with no calls are not necessarily ``leaf
2204 GCC assigns register numbers before it knows whether the function is
2205 suitable for leaf function treatment. So it needs to renumber the
2206 registers in order to output a leaf function. The following macros
2209 @defmac LEAF_REGISTERS
2210 Name of a char vector, indexed by hard register number, which
2211 contains 1 for a register that is allowable in a candidate for leaf
2214 If leaf function treatment involves renumbering the registers, then the
2215 registers marked here should be the ones before renumbering---those that
2216 GCC would ordinarily allocate. The registers which will actually be
2217 used in the assembler code, after renumbering, should not be marked with 1
2220 Define this macro only if the target machine offers a way to optimize
2221 the treatment of leaf functions.
2224 @defmac LEAF_REG_REMAP (@var{regno})
2225 A C expression whose value is the register number to which @var{regno}
2226 should be renumbered, when a function is treated as a leaf function.
2228 If @var{regno} is a register number which should not appear in a leaf
2229 function before renumbering, then the expression should yield @minus{}1, which
2230 will cause the compiler to abort.
2232 Define this macro only if the target machine offers a way to optimize the
2233 treatment of leaf functions, and registers need to be renumbered to do
2237 @findex current_function_is_leaf
2238 @findex current_function_uses_only_leaf_regs
2239 @code{TARGET_ASM_FUNCTION_PROLOGUE} and
2240 @code{TARGET_ASM_FUNCTION_EPILOGUE} must usually treat leaf functions
2241 specially. They can test the C variable @code{current_function_is_leaf}
2242 which is nonzero for leaf functions. @code{current_function_is_leaf} is
2243 set prior to local register allocation and is valid for the remaining
2244 compiler passes. They can also test the C variable
2245 @code{current_function_uses_only_leaf_regs} which is nonzero for leaf
2246 functions which only use leaf registers.
2247 @code{current_function_uses_only_leaf_regs} is valid after all passes
2248 that modify the instructions have been run and is only useful if
2249 @code{LEAF_REGISTERS} is defined.
2250 @c changed this to fix overfull. ALSO: why the "it" at the beginning
2251 @c of the next paragraph?! --mew 2feb93
2253 @node Stack Registers
2254 @subsection Registers That Form a Stack
2256 There are special features to handle computers where some of the
2257 ``registers'' form a stack. Stack registers are normally written by
2258 pushing onto the stack, and are numbered relative to the top of the
2261 Currently, GCC can only handle one group of stack-like registers, and
2262 they must be consecutively numbered. Furthermore, the existing
2263 support for stack-like registers is specific to the 80387 floating
2264 point coprocessor. If you have a new architecture that uses
2265 stack-like registers, you will need to do substantial work on
2266 @file{reg-stack.c} and write your machine description to cooperate
2267 with it, as well as defining these macros.
2270 Define this if the machine has any stack-like registers.
2273 @defmac STACK_REG_COVER_CLASS
2274 This is a cover class containing the stack registers. Define this if
2275 the machine has any stack-like registers.
2278 @defmac FIRST_STACK_REG
2279 The number of the first stack-like register. This one is the top
2283 @defmac LAST_STACK_REG
2284 The number of the last stack-like register. This one is the bottom of
2288 @node Register Classes
2289 @section Register Classes
2290 @cindex register class definitions
2291 @cindex class definitions, register
2293 On many machines, the numbered registers are not all equivalent.
2294 For example, certain registers may not be allowed for indexed addressing;
2295 certain registers may not be allowed in some instructions. These machine
2296 restrictions are described to the compiler using @dfn{register classes}.
2298 You define a number of register classes, giving each one a name and saying
2299 which of the registers belong to it. Then you can specify register classes
2300 that are allowed as operands to particular instruction patterns.
2304 In general, each register will belong to several classes. In fact, one
2305 class must be named @code{ALL_REGS} and contain all the registers. Another
2306 class must be named @code{NO_REGS} and contain no registers. Often the
2307 union of two classes will be another class; however, this is not required.
2309 @findex GENERAL_REGS
2310 One of the classes must be named @code{GENERAL_REGS}. There is nothing
2311 terribly special about the name, but the operand constraint letters
2312 @samp{r} and @samp{g} specify this class. If @code{GENERAL_REGS} is
2313 the same as @code{ALL_REGS}, just define it as a macro which expands
2316 Order the classes so that if class @var{x} is contained in class @var{y}
2317 then @var{x} has a lower class number than @var{y}.
2319 The way classes other than @code{GENERAL_REGS} are specified in operand
2320 constraints is through machine-dependent operand constraint letters.
2321 You can define such letters to correspond to various classes, then use
2322 them in operand constraints.
2324 You must define the narrowest register classes for allocatable
2325 registers, so that each class either has no subclasses, or that for
2326 some mode, the move cost between registers within the class is
2327 cheaper than moving a register in the class to or from memory
2330 You should define a class for the union of two classes whenever some
2331 instruction allows both classes. For example, if an instruction allows
2332 either a floating point (coprocessor) register or a general register for a
2333 certain operand, you should define a class @code{FLOAT_OR_GENERAL_REGS}
2334 which includes both of them. Otherwise you will get suboptimal code,
2335 or even internal compiler errors when reload cannot find a register in the
2336 class computed via @code{reg_class_subunion}.
2338 You must also specify certain redundant information about the register
2339 classes: for each class, which classes contain it and which ones are
2340 contained in it; for each pair of classes, the largest class contained
2343 When a value occupying several consecutive registers is expected in a
2344 certain class, all the registers used must belong to that class.
2345 Therefore, register classes cannot be used to enforce a requirement for
2346 a register pair to start with an even-numbered register. The way to
2347 specify this requirement is with @code{HARD_REGNO_MODE_OK}.
2349 Register classes used for input-operands of bitwise-and or shift
2350 instructions have a special requirement: each such class must have, for
2351 each fixed-point machine mode, a subclass whose registers can transfer that
2352 mode to or from memory. For example, on some machines, the operations for
2353 single-byte values (@code{QImode}) are limited to certain registers. When
2354 this is so, each register class that is used in a bitwise-and or shift
2355 instruction must have a subclass consisting of registers from which
2356 single-byte values can be loaded or stored. This is so that
2357 @code{PREFERRED_RELOAD_CLASS} can always have a possible value to return.
2359 @deftp {Data type} {enum reg_class}
2360 An enumerated type that must be defined with all the register class names
2361 as enumerated values. @code{NO_REGS} must be first. @code{ALL_REGS}
2362 must be the last register class, followed by one more enumerated value,
2363 @code{LIM_REG_CLASSES}, which is not a register class but rather
2364 tells how many classes there are.
2366 Each register class has a number, which is the value of casting
2367 the class name to type @code{int}. The number serves as an index
2368 in many of the tables described below.
2371 @defmac N_REG_CLASSES
2372 The number of distinct register classes, defined as follows:
2375 #define N_REG_CLASSES (int) LIM_REG_CLASSES
2379 @defmac REG_CLASS_NAMES
2380 An initializer containing the names of the register classes as C string
2381 constants. These names are used in writing some of the debugging dumps.
2384 @defmac REG_CLASS_CONTENTS
2385 An initializer containing the contents of the register classes, as integers
2386 which are bit masks. The @var{n}th integer specifies the contents of class
2387 @var{n}. The way the integer @var{mask} is interpreted is that
2388 register @var{r} is in the class if @code{@var{mask} & (1 << @var{r})} is 1.
2390 When the machine has more than 32 registers, an integer does not suffice.
2391 Then the integers are replaced by sub-initializers, braced groupings containing
2392 several integers. Each sub-initializer must be suitable as an initializer
2393 for the type @code{HARD_REG_SET} which is defined in @file{hard-reg-set.h}.
2394 In this situation, the first integer in each sub-initializer corresponds to
2395 registers 0 through 31, the second integer to registers 32 through 63, and
2399 @defmac REGNO_REG_CLASS (@var{regno})
2400 A C expression whose value is a register class containing hard register
2401 @var{regno}. In general there is more than one such class; choose a class
2402 which is @dfn{minimal}, meaning that no smaller class also contains the
2406 @defmac BASE_REG_CLASS
2407 A macro whose definition is the name of the class to which a valid
2408 base register must belong. A base register is one used in an address
2409 which is the register value plus a displacement.
2412 @defmac MODE_BASE_REG_CLASS (@var{mode})
2413 This is a variation of the @code{BASE_REG_CLASS} macro which allows
2414 the selection of a base register in a mode dependent manner. If
2415 @var{mode} is VOIDmode then it should return the same value as
2416 @code{BASE_REG_CLASS}.
2419 @defmac MODE_BASE_REG_REG_CLASS (@var{mode})
2420 A C expression whose value is the register class to which a valid
2421 base register must belong in order to be used in a base plus index
2422 register address. You should define this macro if base plus index
2423 addresses have different requirements than other base register uses.
2426 @defmac MODE_CODE_BASE_REG_CLASS (@var{mode}, @var{outer_code}, @var{index_code})
2427 A C expression whose value is the register class to which a valid
2428 base register must belong. @var{outer_code} and @var{index_code} define the
2429 context in which the base register occurs. @var{outer_code} is the code of
2430 the immediately enclosing expression (@code{MEM} for the top level of an
2431 address, @code{ADDRESS} for something that occurs in an
2432 @code{address_operand}). @var{index_code} is the code of the corresponding
2433 index expression if @var{outer_code} is @code{PLUS}; @code{SCRATCH} otherwise.
2436 @defmac INDEX_REG_CLASS
2437 A macro whose definition is the name of the class to which a valid
2438 index register must belong. An index register is one used in an
2439 address where its value is either multiplied by a scale factor or
2440 added to another register (as well as added to a displacement).
2443 @defmac REGNO_OK_FOR_BASE_P (@var{num})
2444 A C expression which is nonzero if register number @var{num} is
2445 suitable for use as a base register in operand addresses.
2448 @defmac REGNO_MODE_OK_FOR_BASE_P (@var{num}, @var{mode})
2449 A C expression that is just like @code{REGNO_OK_FOR_BASE_P}, except that
2450 that expression may examine the mode of the memory reference in
2451 @var{mode}. You should define this macro if the mode of the memory
2452 reference affects whether a register may be used as a base register. If
2453 you define this macro, the compiler will use it instead of
2454 @code{REGNO_OK_FOR_BASE_P}. The mode may be @code{VOIDmode} for
2455 addresses that appear outside a @code{MEM}, i.e., as an
2456 @code{address_operand}.
2459 @defmac REGNO_MODE_OK_FOR_REG_BASE_P (@var{num}, @var{mode})
2460 A C expression which is nonzero if register number @var{num} is suitable for
2461 use as a base register in base plus index operand addresses, accessing
2462 memory in mode @var{mode}. It may be either a suitable hard register or a
2463 pseudo register that has been allocated such a hard register. You should
2464 define this macro if base plus index addresses have different requirements
2465 than other base register uses.
2467 Use of this macro is deprecated; please use the more general
2468 @code{REGNO_MODE_CODE_OK_FOR_BASE_P}.
2471 @defmac REGNO_MODE_CODE_OK_FOR_BASE_P (@var{num}, @var{mode}, @var{outer_code}, @var{index_code})
2472 A C expression that is just like @code{REGNO_MODE_OK_FOR_BASE_P}, except
2473 that that expression may examine the context in which the register
2474 appears in the memory reference. @var{outer_code} is the code of the
2475 immediately enclosing expression (@code{MEM} if at the top level of the
2476 address, @code{ADDRESS} for something that occurs in an
2477 @code{address_operand}). @var{index_code} is the code of the
2478 corresponding index expression if @var{outer_code} is @code{PLUS};
2479 @code{SCRATCH} otherwise. The mode may be @code{VOIDmode} for addresses
2480 that appear outside a @code{MEM}, i.e., as an @code{address_operand}.
2483 @defmac REGNO_OK_FOR_INDEX_P (@var{num})
2484 A C expression which is nonzero if register number @var{num} is
2485 suitable for use as an index register in operand addresses. It may be
2486 either a suitable hard register or a pseudo register that has been
2487 allocated such a hard register.
2489 The difference between an index register and a base register is that
2490 the index register may be scaled. If an address involves the sum of
2491 two registers, neither one of them scaled, then either one may be
2492 labeled the ``base'' and the other the ``index''; but whichever
2493 labeling is used must fit the machine's constraints of which registers
2494 may serve in each capacity. The compiler will try both labelings,
2495 looking for one that is valid, and will reload one or both registers
2496 only if neither labeling works.
2499 @hook TARGET_PREFERRED_RENAME_CLASS
2501 @hook TARGET_PREFERRED_RELOAD_CLASS
2502 A target hook that places additional restrictions on the register class
2503 to use when it is necessary to copy value @var{x} into a register in class
2504 @var{rclass}. The value is a register class; perhaps @var{rclass}, or perhaps
2505 another, smaller class.
2507 The default version of this hook always returns value of @code{rclass} argument.
2509 Sometimes returning a more restrictive class makes better code. For
2510 example, on the 68000, when @var{x} is an integer constant that is in range
2511 for a @samp{moveq} instruction, the value of this macro is always
2512 @code{DATA_REGS} as long as @var{rclass} includes the data registers.
2513 Requiring a data register guarantees that a @samp{moveq} will be used.
2515 One case where @code{TARGET_PREFERRED_RELOAD_CLASS} must not return
2516 @var{rclass} is if @var{x} is a legitimate constant which cannot be
2517 loaded into some register class. By returning @code{NO_REGS} you can
2518 force @var{x} into a memory location. For example, rs6000 can load
2519 immediate values into general-purpose registers, but does not have an
2520 instruction for loading an immediate value into a floating-point
2521 register, so @code{TARGET_PREFERRED_RELOAD_CLASS} returns @code{NO_REGS} when
2522 @var{x} is a floating-point constant. If the constant can't be loaded
2523 into any kind of register, code generation will be better if
2524 @code{TARGET_LEGITIMATE_CONSTANT_P} makes the constant illegitimate instead
2525 of using @code{TARGET_PREFERRED_RELOAD_CLASS}.
2527 If an insn has pseudos in it after register allocation, reload will go
2528 through the alternatives and call repeatedly @code{TARGET_PREFERRED_RELOAD_CLASS}
2529 to find the best one. Returning @code{NO_REGS}, in this case, makes
2530 reload add a @code{!} in front of the constraint: the x86 back-end uses
2531 this feature to discourage usage of 387 registers when math is done in
2532 the SSE registers (and vice versa).
2535 @defmac PREFERRED_RELOAD_CLASS (@var{x}, @var{class})
2536 A C expression that places additional restrictions on the register class
2537 to use when it is necessary to copy value @var{x} into a register in class
2538 @var{class}. The value is a register class; perhaps @var{class}, or perhaps
2539 another, smaller class. On many machines, the following definition is
2543 #define PREFERRED_RELOAD_CLASS(X,CLASS) CLASS
2546 Sometimes returning a more restrictive class makes better code. For
2547 example, on the 68000, when @var{x} is an integer constant that is in range
2548 for a @samp{moveq} instruction, the value of this macro is always
2549 @code{DATA_REGS} as long as @var{class} includes the data registers.
2550 Requiring a data register guarantees that a @samp{moveq} will be used.
2552 One case where @code{PREFERRED_RELOAD_CLASS} must not return
2553 @var{class} is if @var{x} is a legitimate constant which cannot be
2554 loaded into some register class. By returning @code{NO_REGS} you can
2555 force @var{x} into a memory location. For example, rs6000 can load
2556 immediate values into general-purpose registers, but does not have an
2557 instruction for loading an immediate value into a floating-point
2558 register, so @code{PREFERRED_RELOAD_CLASS} returns @code{NO_REGS} when
2559 @var{x} is a floating-point constant. If the constant can't be loaded
2560 into any kind of register, code generation will be better if
2561 @code{TARGET_LEGITIMATE_CONSTANT_P} makes the constant illegitimate instead
2562 of using @code{TARGET_PREFERRED_RELOAD_CLASS}.
2564 If an insn has pseudos in it after register allocation, reload will go
2565 through the alternatives and call repeatedly @code{PREFERRED_RELOAD_CLASS}
2566 to find the best one. Returning @code{NO_REGS}, in this case, makes
2567 reload add a @code{!} in front of the constraint: the x86 back-end uses
2568 this feature to discourage usage of 387 registers when math is done in
2569 the SSE registers (and vice versa).
2572 @hook TARGET_PREFERRED_OUTPUT_RELOAD_CLASS
2573 Like @code{TARGET_PREFERRED_RELOAD_CLASS}, but for output reloads instead of
2576 The default version of this hook always returns value of @code{rclass}
2579 You can also use @code{TARGET_PREFERRED_OUTPUT_RELOAD_CLASS} to discourage
2580 reload from using some alternatives, like @code{TARGET_PREFERRED_RELOAD_CLASS}.
2583 @defmac LIMIT_RELOAD_CLASS (@var{mode}, @var{class})
2584 A C expression that places additional restrictions on the register class
2585 to use when it is necessary to be able to hold a value of mode
2586 @var{mode} in a reload register for which class @var{class} would
2589 Unlike @code{PREFERRED_RELOAD_CLASS}, this macro should be used when
2590 there are certain modes that simply can't go in certain reload classes.
2592 The value is a register class; perhaps @var{class}, or perhaps another,
2595 Don't define this macro unless the target machine has limitations which
2596 require the macro to do something nontrivial.
2599 @hook TARGET_SECONDARY_RELOAD
2600 Many machines have some registers that cannot be copied directly to or
2601 from memory or even from other types of registers. An example is the
2602 @samp{MQ} register, which on most machines, can only be copied to or
2603 from general registers, but not memory. Below, we shall be using the
2604 term 'intermediate register' when a move operation cannot be performed
2605 directly, but has to be done by copying the source into the intermediate
2606 register first, and then copying the intermediate register to the
2607 destination. An intermediate register always has the same mode as
2608 source and destination. Since it holds the actual value being copied,
2609 reload might apply optimizations to re-use an intermediate register
2610 and eliding the copy from the source when it can determine that the
2611 intermediate register still holds the required value.
2613 Another kind of secondary reload is required on some machines which
2614 allow copying all registers to and from memory, but require a scratch
2615 register for stores to some memory locations (e.g., those with symbolic
2616 address on the RT, and those with certain symbolic address on the SPARC
2617 when compiling PIC)@. Scratch registers need not have the same mode
2618 as the value being copied, and usually hold a different value than
2619 that being copied. Special patterns in the md file are needed to
2620 describe how the copy is performed with the help of the scratch register;
2621 these patterns also describe the number, register class(es) and mode(s)
2622 of the scratch register(s).
2624 In some cases, both an intermediate and a scratch register are required.
2626 For input reloads, this target hook is called with nonzero @var{in_p},
2627 and @var{x} is an rtx that needs to be copied to a register of class
2628 @var{reload_class} in @var{reload_mode}. For output reloads, this target
2629 hook is called with zero @var{in_p}, and a register of class @var{reload_class}
2630 needs to be copied to rtx @var{x} in @var{reload_mode}.
2632 If copying a register of @var{reload_class} from/to @var{x} requires
2633 an intermediate register, the hook @code{secondary_reload} should
2634 return the register class required for this intermediate register.
2635 If no intermediate register is required, it should return NO_REGS.
2636 If more than one intermediate register is required, describe the one
2637 that is closest in the copy chain to the reload register.
2639 If scratch registers are needed, you also have to describe how to
2640 perform the copy from/to the reload register to/from this
2641 closest intermediate register. Or if no intermediate register is
2642 required, but still a scratch register is needed, describe the
2643 copy from/to the reload register to/from the reload operand @var{x}.
2645 You do this by setting @code{sri->icode} to the instruction code of a pattern
2646 in the md file which performs the move. Operands 0 and 1 are the output
2647 and input of this copy, respectively. Operands from operand 2 onward are
2648 for scratch operands. These scratch operands must have a mode, and a
2649 single-register-class
2650 @c [later: or memory]
2653 When an intermediate register is used, the @code{secondary_reload}
2654 hook will be called again to determine how to copy the intermediate
2655 register to/from the reload operand @var{x}, so your hook must also
2656 have code to handle the register class of the intermediate operand.
2658 @c [For later: maybe we'll allow multi-alternative reload patterns -
2659 @c the port maintainer could name a mov<mode> pattern that has clobbers -
2660 @c and match the constraints of input and output to determine the required
2661 @c alternative. A restriction would be that constraints used to match
2662 @c against reloads registers would have to be written as register class
2663 @c constraints, or we need a new target macro / hook that tells us if an
2664 @c arbitrary constraint can match an unknown register of a given class.
2665 @c Such a macro / hook would also be useful in other places.]
2668 @var{x} might be a pseudo-register or a @code{subreg} of a
2669 pseudo-register, which could either be in a hard register or in memory.
2670 Use @code{true_regnum} to find out; it will return @minus{}1 if the pseudo is
2671 in memory and the hard register number if it is in a register.
2673 Scratch operands in memory (constraint @code{"=m"} / @code{"=&m"}) are
2674 currently not supported. For the time being, you will have to continue
2675 to use @code{SECONDARY_MEMORY_NEEDED} for that purpose.
2677 @code{copy_cost} also uses this target hook to find out how values are
2678 copied. If you want it to include some extra cost for the need to allocate
2679 (a) scratch register(s), set @code{sri->extra_cost} to the additional cost.
2680 Or if two dependent moves are supposed to have a lower cost than the sum
2681 of the individual moves due to expected fortuitous scheduling and/or special
2682 forwarding logic, you can set @code{sri->extra_cost} to a negative amount.
2685 @defmac SECONDARY_RELOAD_CLASS (@var{class}, @var{mode}, @var{x})
2686 @defmacx SECONDARY_INPUT_RELOAD_CLASS (@var{class}, @var{mode}, @var{x})
2687 @defmacx SECONDARY_OUTPUT_RELOAD_CLASS (@var{class}, @var{mode}, @var{x})
2688 These macros are obsolete, new ports should use the target hook
2689 @code{TARGET_SECONDARY_RELOAD} instead.
2691 These are obsolete macros, replaced by the @code{TARGET_SECONDARY_RELOAD}
2692 target hook. Older ports still define these macros to indicate to the
2693 reload phase that it may
2694 need to allocate at least one register for a reload in addition to the
2695 register to contain the data. Specifically, if copying @var{x} to a
2696 register @var{class} in @var{mode} requires an intermediate register,
2697 you were supposed to define @code{SECONDARY_INPUT_RELOAD_CLASS} to return the
2698 largest register class all of whose registers can be used as
2699 intermediate registers or scratch registers.
2701 If copying a register @var{class} in @var{mode} to @var{x} requires an
2702 intermediate or scratch register, @code{SECONDARY_OUTPUT_RELOAD_CLASS}
2703 was supposed to be defined be defined to return the largest register
2704 class required. If the
2705 requirements for input and output reloads were the same, the macro
2706 @code{SECONDARY_RELOAD_CLASS} should have been used instead of defining both
2709 The values returned by these macros are often @code{GENERAL_REGS}.
2710 Return @code{NO_REGS} if no spare register is needed; i.e., if @var{x}
2711 can be directly copied to or from a register of @var{class} in
2712 @var{mode} without requiring a scratch register. Do not define this
2713 macro if it would always return @code{NO_REGS}.
2715 If a scratch register is required (either with or without an
2716 intermediate register), you were supposed to define patterns for
2717 @samp{reload_in@var{m}} or @samp{reload_out@var{m}}, as required
2718 (@pxref{Standard Names}. These patterns, which were normally
2719 implemented with a @code{define_expand}, should be similar to the
2720 @samp{mov@var{m}} patterns, except that operand 2 is the scratch
2723 These patterns need constraints for the reload register and scratch
2725 contain a single register class. If the original reload register (whose
2726 class is @var{class}) can meet the constraint given in the pattern, the
2727 value returned by these macros is used for the class of the scratch
2728 register. Otherwise, two additional reload registers are required.
2729 Their classes are obtained from the constraints in the insn pattern.
2731 @var{x} might be a pseudo-register or a @code{subreg} of a
2732 pseudo-register, which could either be in a hard register or in memory.
2733 Use @code{true_regnum} to find out; it will return @minus{}1 if the pseudo is
2734 in memory and the hard register number if it is in a register.
2736 These macros should not be used in the case where a particular class of
2737 registers can only be copied to memory and not to another class of
2738 registers. In that case, secondary reload registers are not needed and
2739 would not be helpful. Instead, a stack location must be used to perform
2740 the copy and the @code{mov@var{m}} pattern should use memory as an
2741 intermediate storage. This case often occurs between floating-point and
2745 @defmac SECONDARY_MEMORY_NEEDED (@var{class1}, @var{class2}, @var{m})
2746 Certain machines have the property that some registers cannot be copied
2747 to some other registers without using memory. Define this macro on
2748 those machines to be a C expression that is nonzero if objects of mode
2749 @var{m} in registers of @var{class1} can only be copied to registers of
2750 class @var{class2} by storing a register of @var{class1} into memory
2751 and loading that memory location into a register of @var{class2}.
2753 Do not define this macro if its value would always be zero.
2756 @defmac SECONDARY_MEMORY_NEEDED_RTX (@var{mode})
2757 Normally when @code{SECONDARY_MEMORY_NEEDED} is defined, the compiler
2758 allocates a stack slot for a memory location needed for register copies.
2759 If this macro is defined, the compiler instead uses the memory location
2760 defined by this macro.
2762 Do not define this macro if you do not define
2763 @code{SECONDARY_MEMORY_NEEDED}.
2766 @defmac SECONDARY_MEMORY_NEEDED_MODE (@var{mode})
2767 When the compiler needs a secondary memory location to copy between two
2768 registers of mode @var{mode}, it normally allocates sufficient memory to
2769 hold a quantity of @code{BITS_PER_WORD} bits and performs the store and
2770 load operations in a mode that many bits wide and whose class is the
2771 same as that of @var{mode}.
2773 This is right thing to do on most machines because it ensures that all
2774 bits of the register are copied and prevents accesses to the registers
2775 in a narrower mode, which some machines prohibit for floating-point
2778 However, this default behavior is not correct on some machines, such as
2779 the DEC Alpha, that store short integers in floating-point registers
2780 differently than in integer registers. On those machines, the default
2781 widening will not work correctly and you must define this macro to
2782 suppress that widening in some cases. See the file @file{alpha.h} for
2785 Do not define this macro if you do not define
2786 @code{SECONDARY_MEMORY_NEEDED} or if widening @var{mode} to a mode that
2787 is @code{BITS_PER_WORD} bits wide is correct for your machine.
2790 @hook TARGET_CLASS_LIKELY_SPILLED_P
2791 A target hook which returns @code{true} if pseudos that have been assigned
2792 to registers of class @var{rclass} would likely be spilled because
2793 registers of @var{rclass} are needed for spill registers.
2795 The default version of this target hook returns @code{true} if @var{rclass}
2796 has exactly one register and @code{false} otherwise. On most machines, this
2797 default should be used. Only use this target hook to some other expression
2798 if pseudos allocated by @file{local-alloc.c} end up in memory because their
2799 hard registers were needed for spill registers. If this target hook returns
2800 @code{false} for those classes, those pseudos will only be allocated by
2801 @file{global.c}, which knows how to reallocate the pseudo to another
2802 register. If there would not be another register available for reallocation,
2803 you should not change the implementation of this target hook since
2804 the only effect of such implementation would be to slow down register
2808 @hook TARGET_CLASS_MAX_NREGS
2809 A target hook returns the maximum number of consecutive registers
2810 of class @var{rclass} needed to hold a value of mode @var{mode}.
2812 This is closely related to the macro @code{HARD_REGNO_NREGS}. In fact,
2813 the value returned by @code{TARGET_CLASS_MAX_NREGS (@var{rclass},
2814 @var{mode})} target hook should be the maximum value of
2815 @code{HARD_REGNO_NREGS (@var{regno}, @var{mode})} for all @var{regno}
2816 values in the class @var{rclass}.
2818 This target hook helps control the handling of multiple-word values
2821 The default version of this target hook returns the size of @var{mode}
2825 @defmac CLASS_MAX_NREGS (@var{class}, @var{mode})
2826 A C expression for the maximum number of consecutive registers
2827 of class @var{class} needed to hold a value of mode @var{mode}.
2829 This is closely related to the macro @code{HARD_REGNO_NREGS}. In fact,
2830 the value of the macro @code{CLASS_MAX_NREGS (@var{class}, @var{mode})}
2831 should be the maximum value of @code{HARD_REGNO_NREGS (@var{regno},
2832 @var{mode})} for all @var{regno} values in the class @var{class}.
2834 This macro helps control the handling of multiple-word values
2838 @defmac CANNOT_CHANGE_MODE_CLASS (@var{from}, @var{to}, @var{class})
2839 If defined, a C expression that returns nonzero for a @var{class} for which
2840 a change from mode @var{from} to mode @var{to} is invalid.
2842 For the example, loading 32-bit integer or floating-point objects into
2843 floating-point registers on the Alpha extends them to 64 bits.
2844 Therefore loading a 64-bit object and then storing it as a 32-bit object
2845 does not store the low-order 32 bits, as would be the case for a normal
2846 register. Therefore, @file{alpha.h} defines @code{CANNOT_CHANGE_MODE_CLASS}
2850 #define CANNOT_CHANGE_MODE_CLASS(FROM, TO, CLASS) \
2851 (GET_MODE_SIZE (FROM) != GET_MODE_SIZE (TO) \
2852 ? reg_classes_intersect_p (FLOAT_REGS, (CLASS)) : 0)
2856 @node Old Constraints
2857 @section Obsolete Macros for Defining Constraints
2858 @cindex defining constraints, obsolete method
2859 @cindex constraints, defining, obsolete method
2861 Machine-specific constraints can be defined with these macros instead
2862 of the machine description constructs described in @ref{Define
2863 Constraints}. This mechanism is obsolete. New ports should not use
2864 it; old ports should convert to the new mechanism.
2866 @defmac CONSTRAINT_LEN (@var{char}, @var{str})
2867 For the constraint at the start of @var{str}, which starts with the letter
2868 @var{c}, return the length. This allows you to have register class /
2869 constant / extra constraints that are longer than a single letter;
2870 you don't need to define this macro if you can do with single-letter
2871 constraints only. The definition of this macro should use
2872 DEFAULT_CONSTRAINT_LEN for all the characters that you don't want
2873 to handle specially.
2874 There are some sanity checks in genoutput.c that check the constraint lengths
2875 for the md file, so you can also use this macro to help you while you are
2876 transitioning from a byzantine single-letter-constraint scheme: when you
2877 return a negative length for a constraint you want to re-use, genoutput
2878 will complain about every instance where it is used in the md file.
2881 @defmac REG_CLASS_FROM_LETTER (@var{char})
2882 A C expression which defines the machine-dependent operand constraint
2883 letters for register classes. If @var{char} is such a letter, the
2884 value should be the register class corresponding to it. Otherwise,
2885 the value should be @code{NO_REGS}. The register letter @samp{r},
2886 corresponding to class @code{GENERAL_REGS}, will not be passed
2887 to this macro; you do not need to handle it.
2890 @defmac REG_CLASS_FROM_CONSTRAINT (@var{char}, @var{str})
2891 Like @code{REG_CLASS_FROM_LETTER}, but you also get the constraint string
2892 passed in @var{str}, so that you can use suffixes to distinguish between
2896 @defmac CONST_OK_FOR_LETTER_P (@var{value}, @var{c})
2897 A C expression that defines the machine-dependent operand constraint
2898 letters (@samp{I}, @samp{J}, @samp{K}, @dots{} @samp{P}) that specify
2899 particular ranges of integer values. If @var{c} is one of those
2900 letters, the expression should check that @var{value}, an integer, is in
2901 the appropriate range and return 1 if so, 0 otherwise. If @var{c} is
2902 not one of those letters, the value should be 0 regardless of
2906 @defmac CONST_OK_FOR_CONSTRAINT_P (@var{value}, @var{c}, @var{str})
2907 Like @code{CONST_OK_FOR_LETTER_P}, but you also get the constraint
2908 string passed in @var{str}, so that you can use suffixes to distinguish
2909 between different variants.
2912 @defmac CONST_DOUBLE_OK_FOR_LETTER_P (@var{value}, @var{c})
2913 A C expression that defines the machine-dependent operand constraint
2914 letters that specify particular ranges of @code{const_double} values
2915 (@samp{G} or @samp{H}).
2917 If @var{c} is one of those letters, the expression should check that
2918 @var{value}, an RTX of code @code{const_double}, is in the appropriate
2919 range and return 1 if so, 0 otherwise. If @var{c} is not one of those
2920 letters, the value should be 0 regardless of @var{value}.
2922 @code{const_double} is used for all floating-point constants and for
2923 @code{DImode} fixed-point constants. A given letter can accept either
2924 or both kinds of values. It can use @code{GET_MODE} to distinguish
2925 between these kinds.
2928 @defmac CONST_DOUBLE_OK_FOR_CONSTRAINT_P (@var{value}, @var{c}, @var{str})
2929 Like @code{CONST_DOUBLE_OK_FOR_LETTER_P}, but you also get the constraint
2930 string passed in @var{str}, so that you can use suffixes to distinguish
2931 between different variants.
2934 @defmac EXTRA_CONSTRAINT (@var{value}, @var{c})
2935 A C expression that defines the optional machine-dependent constraint
2936 letters that can be used to segregate specific types of operands, usually
2937 memory references, for the target machine. Any letter that is not
2938 elsewhere defined and not matched by @code{REG_CLASS_FROM_LETTER} /
2939 @code{REG_CLASS_FROM_CONSTRAINT}
2940 may be used. Normally this macro will not be defined.
2942 If it is required for a particular target machine, it should return 1
2943 if @var{value} corresponds to the operand type represented by the
2944 constraint letter @var{c}. If @var{c} is not defined as an extra
2945 constraint, the value returned should be 0 regardless of @var{value}.
2947 For example, on the ROMP, load instructions cannot have their output
2948 in r0 if the memory reference contains a symbolic address. Constraint
2949 letter @samp{Q} is defined as representing a memory address that does
2950 @emph{not} contain a symbolic address. An alternative is specified with
2951 a @samp{Q} constraint on the input and @samp{r} on the output. The next
2952 alternative specifies @samp{m} on the input and a register class that
2953 does not include r0 on the output.
2956 @defmac EXTRA_CONSTRAINT_STR (@var{value}, @var{c}, @var{str})
2957 Like @code{EXTRA_CONSTRAINT}, but you also get the constraint string passed
2958 in @var{str}, so that you can use suffixes to distinguish between different
2962 @defmac EXTRA_MEMORY_CONSTRAINT (@var{c}, @var{str})
2963 A C expression that defines the optional machine-dependent constraint
2964 letters, amongst those accepted by @code{EXTRA_CONSTRAINT}, that should
2965 be treated like memory constraints by the reload pass.
2967 It should return 1 if the operand type represented by the constraint
2968 at the start of @var{str}, the first letter of which is the letter @var{c},
2969 comprises a subset of all memory references including
2970 all those whose address is simply a base register. This allows the reload
2971 pass to reload an operand, if it does not directly correspond to the operand
2972 type of @var{c}, by copying its address into a base register.
2974 For example, on the S/390, some instructions do not accept arbitrary
2975 memory references, but only those that do not make use of an index
2976 register. The constraint letter @samp{Q} is defined via
2977 @code{EXTRA_CONSTRAINT} as representing a memory address of this type.
2978 If the letter @samp{Q} is marked as @code{EXTRA_MEMORY_CONSTRAINT},
2979 a @samp{Q} constraint can handle any memory operand, because the
2980 reload pass knows it can be reloaded by copying the memory address
2981 into a base register if required. This is analogous to the way
2982 an @samp{o} constraint can handle any memory operand.
2985 @defmac EXTRA_ADDRESS_CONSTRAINT (@var{c}, @var{str})
2986 A C expression that defines the optional machine-dependent constraint
2987 letters, amongst those accepted by @code{EXTRA_CONSTRAINT} /
2988 @code{EXTRA_CONSTRAINT_STR}, that should
2989 be treated like address constraints by the reload pass.
2991 It should return 1 if the operand type represented by the constraint
2992 at the start of @var{str}, which starts with the letter @var{c}, comprises
2993 a subset of all memory addresses including
2994 all those that consist of just a base register. This allows the reload
2995 pass to reload an operand, if it does not directly correspond to the operand
2996 type of @var{str}, by copying it into a base register.
2998 Any constraint marked as @code{EXTRA_ADDRESS_CONSTRAINT} can only
2999 be used with the @code{address_operand} predicate. It is treated
3000 analogously to the @samp{p} constraint.
3003 @node Stack and Calling
3004 @section Stack Layout and Calling Conventions
3005 @cindex calling conventions
3007 @c prevent bad page break with this line
3008 This describes the stack layout and calling conventions.
3012 * Exception Handling::
3017 * Register Arguments::
3019 * Aggregate Return::
3024 * Stack Smashing Protection::
3028 @subsection Basic Stack Layout
3029 @cindex stack frame layout
3030 @cindex frame layout
3032 @c prevent bad page break with this line
3033 Here is the basic stack layout.
3035 @defmac STACK_GROWS_DOWNWARD
3036 Define this macro if pushing a word onto the stack moves the stack
3037 pointer to a smaller address.
3039 When we say, ``define this macro if @dots{}'', it means that the
3040 compiler checks this macro only with @code{#ifdef} so the precise
3041 definition used does not matter.
3044 @defmac STACK_PUSH_CODE
3045 This macro defines the operation used when something is pushed
3046 on the stack. In RTL, a push operation will be
3047 @code{(set (mem (STACK_PUSH_CODE (reg sp))) @dots{})}
3049 The choices are @code{PRE_DEC}, @code{POST_DEC}, @code{PRE_INC},
3050 and @code{POST_INC}. Which of these is correct depends on
3051 the stack direction and on whether the stack pointer points
3052 to the last item on the stack or whether it points to the
3053 space for the next item on the stack.
3055 The default is @code{PRE_DEC} when @code{STACK_GROWS_DOWNWARD} is
3056 defined, which is almost always right, and @code{PRE_INC} otherwise,
3057 which is often wrong.
3060 @defmac FRAME_GROWS_DOWNWARD
3061 Define this macro to nonzero value if the addresses of local variable slots
3062 are at negative offsets from the frame pointer.
3065 @defmac ARGS_GROW_DOWNWARD
3066 Define this macro if successive arguments to a function occupy decreasing
3067 addresses on the stack.
3070 @defmac STARTING_FRAME_OFFSET
3071 Offset from the frame pointer to the first local variable slot to be allocated.
3073 If @code{FRAME_GROWS_DOWNWARD}, find the next slot's offset by
3074 subtracting the first slot's length from @code{STARTING_FRAME_OFFSET}.
3075 Otherwise, it is found by adding the length of the first slot to the
3076 value @code{STARTING_FRAME_OFFSET}.
3077 @c i'm not sure if the above is still correct.. had to change it to get
3078 @c rid of an overfull. --mew 2feb93
3081 @defmac STACK_ALIGNMENT_NEEDED
3082 Define to zero to disable final alignment of the stack during reload.
3083 The nonzero default for this macro is suitable for most ports.
3085 On ports where @code{STARTING_FRAME_OFFSET} is nonzero or where there
3086 is a register save block following the local block that doesn't require
3087 alignment to @code{STACK_BOUNDARY}, it may be beneficial to disable
3088 stack alignment and do it in the backend.
3091 @defmac STACK_POINTER_OFFSET
3092 Offset from the stack pointer register to the first location at which
3093 outgoing arguments are placed. If not specified, the default value of
3094 zero is used. This is the proper value for most machines.
3096 If @code{ARGS_GROW_DOWNWARD}, this is the offset to the location above
3097 the first location at which outgoing arguments are placed.
3100 @defmac FIRST_PARM_OFFSET (@var{fundecl})
3101 Offset from the argument pointer register to the first argument's
3102 address. On some machines it may depend on the data type of the
3105 If @code{ARGS_GROW_DOWNWARD}, this is the offset to the location above
3106 the first argument's address.
3109 @defmac STACK_DYNAMIC_OFFSET (@var{fundecl})
3110 Offset from the stack pointer register to an item dynamically allocated
3111 on the stack, e.g., by @code{alloca}.
3113 The default value for this macro is @code{STACK_POINTER_OFFSET} plus the
3114 length of the outgoing arguments. The default is correct for most
3115 machines. See @file{function.c} for details.
3118 @defmac INITIAL_FRAME_ADDRESS_RTX
3119 A C expression whose value is RTL representing the address of the initial
3120 stack frame. This address is passed to @code{RETURN_ADDR_RTX} and
3121 @code{DYNAMIC_CHAIN_ADDRESS}. If you don't define this macro, a reasonable
3122 default value will be used. Define this macro in order to make frame pointer
3123 elimination work in the presence of @code{__builtin_frame_address (count)} and
3124 @code{__builtin_return_address (count)} for @code{count} not equal to zero.
3127 @defmac DYNAMIC_CHAIN_ADDRESS (@var{frameaddr})
3128 A C expression whose value is RTL representing the address in a stack
3129 frame where the pointer to the caller's frame is stored. Assume that
3130 @var{frameaddr} is an RTL expression for the address of the stack frame
3133 If you don't define this macro, the default is to return the value
3134 of @var{frameaddr}---that is, the stack frame address is also the
3135 address of the stack word that points to the previous frame.
3138 @defmac SETUP_FRAME_ADDRESSES
3139 If defined, a C expression that produces the machine-specific code to
3140 setup the stack so that arbitrary frames can be accessed. For example,
3141 on the SPARC, we must flush all of the register windows to the stack
3142 before we can access arbitrary stack frames. You will seldom need to
3146 @hook TARGET_BUILTIN_SETJMP_FRAME_VALUE
3147 This target hook should return an rtx that is used to store
3148 the address of the current frame into the built in @code{setjmp} buffer.
3149 The default value, @code{virtual_stack_vars_rtx}, is correct for most
3150 machines. One reason you may need to define this target hook is if
3151 @code{hard_frame_pointer_rtx} is the appropriate value on your machine.
3154 @defmac FRAME_ADDR_RTX (@var{frameaddr})
3155 A C expression whose value is RTL representing the value of the frame
3156 address for the current frame. @var{frameaddr} is the frame pointer
3157 of the current frame. This is used for __builtin_frame_address.
3158 You need only define this macro if the frame address is not the same
3159 as the frame pointer. Most machines do not need to define it.
3162 @defmac RETURN_ADDR_RTX (@var{count}, @var{frameaddr})
3163 A C expression whose value is RTL representing the value of the return
3164 address for the frame @var{count} steps up from the current frame, after
3165 the prologue. @var{frameaddr} is the frame pointer of the @var{count}
3166 frame, or the frame pointer of the @var{count} @minus{} 1 frame if
3167 @code{RETURN_ADDR_IN_PREVIOUS_FRAME} is defined.
3169 The value of the expression must always be the correct address when
3170 @var{count} is zero, but may be @code{NULL_RTX} if there is no way to
3171 determine the return address of other frames.
3174 @defmac RETURN_ADDR_IN_PREVIOUS_FRAME
3175 Define this if the return address of a particular stack frame is accessed
3176 from the frame pointer of the previous stack frame.
3179 @defmac INCOMING_RETURN_ADDR_RTX
3180 A C expression whose value is RTL representing the location of the
3181 incoming return address at the beginning of any function, before the
3182 prologue. This RTL is either a @code{REG}, indicating that the return
3183 value is saved in @samp{REG}, or a @code{MEM} representing a location in
3186 You only need to define this macro if you want to support call frame
3187 debugging information like that provided by DWARF 2.
3189 If this RTL is a @code{REG}, you should also define
3190 @code{DWARF_FRAME_RETURN_COLUMN} to @code{DWARF_FRAME_REGNUM (REGNO)}.
3193 @defmac DWARF_ALT_FRAME_RETURN_COLUMN
3194 A C expression whose value is an integer giving a DWARF 2 column
3195 number that may be used as an alternative return column. The column
3196 must not correspond to any gcc hard register (that is, it must not
3197 be in the range of @code{DWARF_FRAME_REGNUM}).
3199 This macro can be useful if @code{DWARF_FRAME_RETURN_COLUMN} is set to a
3200 general register, but an alternative column needs to be used for signal
3201 frames. Some targets have also used different frame return columns
3205 @defmac DWARF_ZERO_REG
3206 A C expression whose value is an integer giving a DWARF 2 register
3207 number that is considered to always have the value zero. This should
3208 only be defined if the target has an architected zero register, and
3209 someone decided it was a good idea to use that register number to
3210 terminate the stack backtrace. New ports should avoid this.
3213 @hook TARGET_DWARF_HANDLE_FRAME_UNSPEC
3214 This target hook allows the backend to emit frame-related insns that
3215 contain UNSPECs or UNSPEC_VOLATILEs. The DWARF 2 call frame debugging
3216 info engine will invoke it on insns of the form
3218 (set (reg) (unspec [@dots{}] UNSPEC_INDEX))
3222 (set (reg) (unspec_volatile [@dots{}] UNSPECV_INDEX)).
3224 to let the backend emit the call frame instructions. @var{label} is
3225 the CFI label attached to the insn, @var{pattern} is the pattern of
3226 the insn and @var{index} is @code{UNSPEC_INDEX} or @code{UNSPECV_INDEX}.
3229 @defmac INCOMING_FRAME_SP_OFFSET
3230 A C expression whose value is an integer giving the offset, in bytes,
3231 from the value of the stack pointer register to the top of the stack
3232 frame at the beginning of any function, before the prologue. The top of
3233 the frame is defined to be the value of the stack pointer in the
3234 previous frame, just before the call instruction.
3236 You only need to define this macro if you want to support call frame
3237 debugging information like that provided by DWARF 2.
3240 @defmac ARG_POINTER_CFA_OFFSET (@var{fundecl})
3241 A C expression whose value is an integer giving the offset, in bytes,
3242 from the argument pointer to the canonical frame address (cfa). The
3243 final value should coincide with that calculated by
3244 @code{INCOMING_FRAME_SP_OFFSET}. Which is unfortunately not usable
3245 during virtual register instantiation.
3247 The default value for this macro is
3248 @code{FIRST_PARM_OFFSET (fundecl) + crtl->args.pretend_args_size},
3249 which is correct for most machines; in general, the arguments are found
3250 immediately before the stack frame. Note that this is not the case on
3251 some targets that save registers into the caller's frame, such as SPARC
3252 and rs6000, and so such targets need to define this macro.
3254 You only need to define this macro if the default is incorrect, and you
3255 want to support call frame debugging information like that provided by
3259 @defmac FRAME_POINTER_CFA_OFFSET (@var{fundecl})
3260 If defined, a C expression whose value is an integer giving the offset
3261 in bytes from the frame pointer to the canonical frame address (cfa).
3262 The final value should coincide with that calculated by
3263 @code{INCOMING_FRAME_SP_OFFSET}.
3265 Normally the CFA is calculated as an offset from the argument pointer,
3266 via @code{ARG_POINTER_CFA_OFFSET}, but if the argument pointer is
3267 variable due to the ABI, this may not be possible. If this macro is
3268 defined, it implies that the virtual register instantiation should be
3269 based on the frame pointer instead of the argument pointer. Only one
3270 of @code{FRAME_POINTER_CFA_OFFSET} and @code{ARG_POINTER_CFA_OFFSET}
3274 @defmac CFA_FRAME_BASE_OFFSET (@var{fundecl})
3275 If defined, a C expression whose value is an integer giving the offset
3276 in bytes from the canonical frame address (cfa) to the frame base used
3277 in DWARF 2 debug information. The default is zero. A different value
3278 may reduce the size of debug information on some ports.
3281 @node Exception Handling
3282 @subsection Exception Handling Support
3283 @cindex exception handling
3285 @defmac EH_RETURN_DATA_REGNO (@var{N})
3286 A C expression whose value is the @var{N}th register number used for
3287 data by exception handlers, or @code{INVALID_REGNUM} if fewer than
3288 @var{N} registers are usable.
3290 The exception handling library routines communicate with the exception
3291 handlers via a set of agreed upon registers. Ideally these registers
3292 should be call-clobbered; it is possible to use call-saved registers,
3293 but may negatively impact code size. The target must support at least
3294 2 data registers, but should define 4 if there are enough free registers.
3296 You must define this macro if you want to support call frame exception
3297 handling like that provided by DWARF 2.
3300 @defmac EH_RETURN_STACKADJ_RTX
3301 A C expression whose value is RTL representing a location in which
3302 to store a stack adjustment to be applied before function return.
3303 This is used to unwind the stack to an exception handler's call frame.
3304 It will be assigned zero on code paths that return normally.
3306 Typically this is a call-clobbered hard register that is otherwise
3307 untouched by the epilogue, but could also be a stack slot.
3309 Do not define this macro if the stack pointer is saved and restored
3310 by the regular prolog and epilog code in the call frame itself; in
3311 this case, the exception handling library routines will update the
3312 stack location to be restored in place. Otherwise, you must define
3313 this macro if you want to support call frame exception handling like
3314 that provided by DWARF 2.
3317 @defmac EH_RETURN_HANDLER_RTX
3318 A C expression whose value is RTL representing a location in which
3319 to store the address of an exception handler to which we should
3320 return. It will not be assigned on code paths that return normally.
3322 Typically this is the location in the call frame at which the normal
3323 return address is stored. For targets that return by popping an
3324 address off the stack, this might be a memory address just below
3325 the @emph{target} call frame rather than inside the current call
3326 frame. If defined, @code{EH_RETURN_STACKADJ_RTX} will have already
3327 been assigned, so it may be used to calculate the location of the
3330 Some targets have more complex requirements than storing to an
3331 address calculable during initial code generation. In that case
3332 the @code{eh_return} instruction pattern should be used instead.
3334 If you want to support call frame exception handling, you must
3335 define either this macro or the @code{eh_return} instruction pattern.
3338 @defmac RETURN_ADDR_OFFSET
3339 If defined, an integer-valued C expression for which rtl will be generated
3340 to add it to the exception handler address before it is searched in the
3341 exception handling tables, and to subtract it again from the address before
3342 using it to return to the exception handler.
3345 @defmac ASM_PREFERRED_EH_DATA_FORMAT (@var{code}, @var{global})
3346 This macro chooses the encoding of pointers embedded in the exception
3347 handling sections. If at all possible, this should be defined such
3348 that the exception handling section will not require dynamic relocations,
3349 and so may be read-only.
3351 @var{code} is 0 for data, 1 for code labels, 2 for function pointers.
3352 @var{global} is true if the symbol may be affected by dynamic relocations.
3353 The macro should return a combination of the @code{DW_EH_PE_*} defines
3354 as found in @file{dwarf2.h}.
3356 If this macro is not defined, pointers will not be encoded but
3357 represented directly.
3360 @defmac ASM_MAYBE_OUTPUT_ENCODED_ADDR_RTX (@var{file}, @var{encoding}, @var{size}, @var{addr}, @var{done})
3361 This macro allows the target to emit whatever special magic is required
3362 to represent the encoding chosen by @code{ASM_PREFERRED_EH_DATA_FORMAT}.
3363 Generic code takes care of pc-relative and indirect encodings; this must
3364 be defined if the target uses text-relative or data-relative encodings.
3366 This is a C statement that branches to @var{done} if the format was
3367 handled. @var{encoding} is the format chosen, @var{size} is the number
3368 of bytes that the format occupies, @var{addr} is the @code{SYMBOL_REF}
3372 @defmac MD_FALLBACK_FRAME_STATE_FOR (@var{context}, @var{fs})
3373 This macro allows the target to add CPU and operating system specific
3374 code to the call-frame unwinder for use when there is no unwind data
3375 available. The most common reason to implement this macro is to unwind
3376 through signal frames.
3378 This macro is called from @code{uw_frame_state_for} in
3379 @file{unwind-dw2.c}, @file{unwind-dw2-xtensa.c} and
3380 @file{unwind-ia64.c}. @var{context} is an @code{_Unwind_Context};
3381 @var{fs} is an @code{_Unwind_FrameState}. Examine @code{context->ra}
3382 for the address of the code being executed and @code{context->cfa} for
3383 the stack pointer value. If the frame can be decoded, the register
3384 save addresses should be updated in @var{fs} and the macro should
3385 evaluate to @code{_URC_NO_REASON}. If the frame cannot be decoded,
3386 the macro should evaluate to @code{_URC_END_OF_STACK}.
3388 For proper signal handling in Java this macro is accompanied by
3389 @code{MAKE_THROW_FRAME}, defined in @file{libjava/include/*-signal.h} headers.
3392 @defmac MD_HANDLE_UNWABI (@var{context}, @var{fs})
3393 This macro allows the target to add operating system specific code to the
3394 call-frame unwinder to handle the IA-64 @code{.unwabi} unwinding directive,
3395 usually used for signal or interrupt frames.
3397 This macro is called from @code{uw_update_context} in @file{unwind-ia64.c}.
3398 @var{context} is an @code{_Unwind_Context};
3399 @var{fs} is an @code{_Unwind_FrameState}. Examine @code{fs->unwabi}
3400 for the abi and context in the @code{.unwabi} directive. If the
3401 @code{.unwabi} directive can be handled, the register save addresses should
3402 be updated in @var{fs}.
3405 @defmac TARGET_USES_WEAK_UNWIND_INFO
3406 A C expression that evaluates to true if the target requires unwind
3407 info to be given comdat linkage. Define it to be @code{1} if comdat
3408 linkage is necessary. The default is @code{0}.
3411 @node Stack Checking
3412 @subsection Specifying How Stack Checking is Done
3414 GCC will check that stack references are within the boundaries of the
3415 stack, if the option @option{-fstack-check} is specified, in one of
3420 If the value of the @code{STACK_CHECK_BUILTIN} macro is nonzero, GCC
3421 will assume that you have arranged for full stack checking to be done
3422 at appropriate places in the configuration files. GCC will not do
3423 other special processing.
3426 If @code{STACK_CHECK_BUILTIN} is zero and the value of the
3427 @code{STACK_CHECK_STATIC_BUILTIN} macro is nonzero, GCC will assume
3428 that you have arranged for static stack checking (checking of the
3429 static stack frame of functions) to be done at appropriate places
3430 in the configuration files. GCC will only emit code to do dynamic
3431 stack checking (checking on dynamic stack allocations) using the third
3435 If neither of the above are true, GCC will generate code to periodically
3436 ``probe'' the stack pointer using the values of the macros defined below.
3439 If neither STACK_CHECK_BUILTIN nor STACK_CHECK_STATIC_BUILTIN is defined,
3440 GCC will change its allocation strategy for large objects if the option
3441 @option{-fstack-check} is specified: they will always be allocated
3442 dynamically if their size exceeds @code{STACK_CHECK_MAX_VAR_SIZE} bytes.
3444 @defmac STACK_CHECK_BUILTIN
3445 A nonzero value if stack checking is done by the configuration files in a
3446 machine-dependent manner. You should define this macro if stack checking
3447 is required by the ABI of your machine or if you would like to do stack
3448 checking in some more efficient way than the generic approach. The default
3449 value of this macro is zero.
3452 @defmac STACK_CHECK_STATIC_BUILTIN
3453 A nonzero value if static stack checking is done by the configuration files
3454 in a machine-dependent manner. You should define this macro if you would
3455 like to do static stack checking in some more efficient way than the generic
3456 approach. The default value of this macro is zero.
3459 @defmac STACK_CHECK_PROBE_INTERVAL_EXP
3460 An integer specifying the interval at which GCC must generate stack probe
3461 instructions, defined as 2 raised to this integer. You will normally
3462 define this macro so that the interval be no larger than the size of
3463 the ``guard pages'' at the end of a stack area. The default value
3464 of 12 (4096-byte interval) is suitable for most systems.
3467 @defmac STACK_CHECK_MOVING_SP
3468 An integer which is nonzero if GCC should move the stack pointer page by page
3469 when doing probes. This can be necessary on systems where the stack pointer
3470 contains the bottom address of the memory area accessible to the executing
3471 thread at any point in time. In this situation an alternate signal stack
3472 is required in order to be able to recover from a stack overflow. The
3473 default value of this macro is zero.
3476 @defmac STACK_CHECK_PROTECT
3477 The number of bytes of stack needed to recover from a stack overflow, for
3478 languages where such a recovery is supported. The default value of 75 words
3479 with the @code{setjmp}/@code{longjmp}-based exception handling mechanism and
3480 8192 bytes with other exception handling mechanisms should be adequate for
3484 The following macros are relevant only if neither STACK_CHECK_BUILTIN
3485 nor STACK_CHECK_STATIC_BUILTIN is defined; you can omit them altogether
3486 in the opposite case.
3488 @defmac STACK_CHECK_MAX_FRAME_SIZE
3489 The maximum size of a stack frame, in bytes. GCC will generate probe
3490 instructions in non-leaf functions to ensure at least this many bytes of
3491 stack are available. If a stack frame is larger than this size, stack
3492 checking will not be reliable and GCC will issue a warning. The
3493 default is chosen so that GCC only generates one instruction on most
3494 systems. You should normally not change the default value of this macro.
3497 @defmac STACK_CHECK_FIXED_FRAME_SIZE
3498 GCC uses this value to generate the above warning message. It
3499 represents the amount of fixed frame used by a function, not including
3500 space for any callee-saved registers, temporaries and user variables.
3501 You need only specify an upper bound for this amount and will normally
3502 use the default of four words.
3505 @defmac STACK_CHECK_MAX_VAR_SIZE
3506 The maximum size, in bytes, of an object that GCC will place in the
3507 fixed area of the stack frame when the user specifies
3508 @option{-fstack-check}.
3509 GCC computed the default from the values of the above macros and you will
3510 normally not need to override that default.
3514 @node Frame Registers
3515 @subsection Registers That Address the Stack Frame
3517 @c prevent bad page break with this line
3518 This discusses registers that address the stack frame.
3520 @defmac STACK_POINTER_REGNUM
3521 The register number of the stack pointer register, which must also be a
3522 fixed register according to @code{FIXED_REGISTERS}. On most machines,
3523 the hardware determines which register this is.
3526 @defmac FRAME_POINTER_REGNUM
3527 The register number of the frame pointer register, which is used to
3528 access automatic variables in the stack frame. On some machines, the
3529 hardware determines which register this is. On other machines, you can
3530 choose any register you wish for this purpose.
3533 @defmac HARD_FRAME_POINTER_REGNUM
3534 On some machines the offset between the frame pointer and starting
3535 offset of the automatic variables is not known until after register
3536 allocation has been done (for example, because the saved registers are
3537 between these two locations). On those machines, define
3538 @code{FRAME_POINTER_REGNUM} the number of a special, fixed register to
3539 be used internally until the offset is known, and define
3540 @code{HARD_FRAME_POINTER_REGNUM} to be the actual hard register number
3541 used for the frame pointer.
3543 You should define this macro only in the very rare circumstances when it
3544 is not possible to calculate the offset between the frame pointer and
3545 the automatic variables until after register allocation has been
3546 completed. When this macro is defined, you must also indicate in your
3547 definition of @code{ELIMINABLE_REGS} how to eliminate
3548 @code{FRAME_POINTER_REGNUM} into either @code{HARD_FRAME_POINTER_REGNUM}
3549 or @code{STACK_POINTER_REGNUM}.
3551 Do not define this macro if it would be the same as
3552 @code{FRAME_POINTER_REGNUM}.
3555 @defmac ARG_POINTER_REGNUM
3556 The register number of the arg pointer register, which is used to access
3557 the function's argument list. On some machines, this is the same as the
3558 frame pointer register. On some machines, the hardware determines which
3559 register this is. On other machines, you can choose any register you
3560 wish for this purpose. If this is not the same register as the frame
3561 pointer register, then you must mark it as a fixed register according to
3562 @code{FIXED_REGISTERS}, or arrange to be able to eliminate it
3563 (@pxref{Elimination}).
3566 @defmac HARD_FRAME_POINTER_IS_FRAME_POINTER
3567 Define this to a preprocessor constant that is nonzero if
3568 @code{hard_frame_pointer_rtx} and @code{frame_pointer_rtx} should be
3569 the same. The default definition is @samp{(HARD_FRAME_POINTER_REGNUM
3570 == FRAME_POINTER_REGNUM)}; you only need to define this macro if that
3571 definition is not suitable for use in preprocessor conditionals.
3574 @defmac HARD_FRAME_POINTER_IS_ARG_POINTER
3575 Define this to a preprocessor constant that is nonzero if
3576 @code{hard_frame_pointer_rtx} and @code{arg_pointer_rtx} should be the
3577 same. The default definition is @samp{(HARD_FRAME_POINTER_REGNUM ==
3578 ARG_POINTER_REGNUM)}; you only need to define this macro if that
3579 definition is not suitable for use in preprocessor conditionals.
3582 @defmac RETURN_ADDRESS_POINTER_REGNUM
3583 The register number of the return address pointer register, which is used to
3584 access the current function's return address from the stack. On some
3585 machines, the return address is not at a fixed offset from the frame
3586 pointer or stack pointer or argument pointer. This register can be defined
3587 to point to the return address on the stack, and then be converted by
3588 @code{ELIMINABLE_REGS} into either the frame pointer or stack pointer.
3590 Do not define this macro unless there is no other way to get the return
3591 address from the stack.
3594 @defmac STATIC_CHAIN_REGNUM
3595 @defmacx STATIC_CHAIN_INCOMING_REGNUM
3596 Register numbers used for passing a function's static chain pointer. If
3597 register windows are used, the register number as seen by the called
3598 function is @code{STATIC_CHAIN_INCOMING_REGNUM}, while the register
3599 number as seen by the calling function is @code{STATIC_CHAIN_REGNUM}. If
3600 these registers are the same, @code{STATIC_CHAIN_INCOMING_REGNUM} need
3603 The static chain register need not be a fixed register.
3605 If the static chain is passed in memory, these macros should not be
3606 defined; instead, the @code{TARGET_STATIC_CHAIN} hook should be used.
3609 @hook TARGET_STATIC_CHAIN
3610 This hook replaces the use of @code{STATIC_CHAIN_REGNUM} et al for
3611 targets that may use different static chain locations for different
3612 nested functions. This may be required if the target has function
3613 attributes that affect the calling conventions of the function and
3614 those calling conventions use different static chain locations.
3616 The default version of this hook uses @code{STATIC_CHAIN_REGNUM} et al.
3618 If the static chain is passed in memory, this hook should be used to
3619 provide rtx giving @code{mem} expressions that denote where they are stored.
3620 Often the @code{mem} expression as seen by the caller will be at an offset
3621 from the stack pointer and the @code{mem} expression as seen by the callee
3622 will be at an offset from the frame pointer.
3623 @findex stack_pointer_rtx
3624 @findex frame_pointer_rtx
3625 @findex arg_pointer_rtx
3626 The variables @code{stack_pointer_rtx}, @code{frame_pointer_rtx}, and
3627 @code{arg_pointer_rtx} will have been initialized and should be used
3628 to refer to those items.
3631 @defmac DWARF_FRAME_REGISTERS
3632 This macro specifies the maximum number of hard registers that can be
3633 saved in a call frame. This is used to size data structures used in
3634 DWARF2 exception handling.
3636 Prior to GCC 3.0, this macro was needed in order to establish a stable
3637 exception handling ABI in the face of adding new hard registers for ISA
3638 extensions. In GCC 3.0 and later, the EH ABI is insulated from changes
3639 in the number of hard registers. Nevertheless, this macro can still be
3640 used to reduce the runtime memory requirements of the exception handling
3641 routines, which can be substantial if the ISA contains a lot of
3642 registers that are not call-saved.
3644 If this macro is not defined, it defaults to
3645 @code{FIRST_PSEUDO_REGISTER}.
3648 @defmac PRE_GCC3_DWARF_FRAME_REGISTERS
3650 This macro is similar to @code{DWARF_FRAME_REGISTERS}, but is provided
3651 for backward compatibility in pre GCC 3.0 compiled code.
3653 If this macro is not defined, it defaults to
3654 @code{DWARF_FRAME_REGISTERS}.
3657 @defmac DWARF_REG_TO_UNWIND_COLUMN (@var{regno})
3659 Define this macro if the target's representation for dwarf registers
3660 is different than the internal representation for unwind column.
3661 Given a dwarf register, this macro should return the internal unwind
3662 column number to use instead.
3664 See the PowerPC's SPE target for an example.
3667 @defmac DWARF_FRAME_REGNUM (@var{regno})
3669 Define this macro if the target's representation for dwarf registers
3670 used in .eh_frame or .debug_frame is different from that used in other
3671 debug info sections. Given a GCC hard register number, this macro
3672 should return the .eh_frame register number. The default is
3673 @code{DBX_REGISTER_NUMBER (@var{regno})}.
3677 @defmac DWARF2_FRAME_REG_OUT (@var{regno}, @var{for_eh})
3679 Define this macro to map register numbers held in the call frame info
3680 that GCC has collected using @code{DWARF_FRAME_REGNUM} to those that
3681 should be output in .debug_frame (@code{@var{for_eh}} is zero) and
3682 .eh_frame (@code{@var{for_eh}} is nonzero). The default is to
3683 return @code{@var{regno}}.
3687 @defmac REG_VALUE_IN_UNWIND_CONTEXT
3689 Define this macro if the target stores register values as
3690 @code{_Unwind_Word} type in unwind context. It should be defined if
3691 target register size is larger than the size of @code{void *}. The
3692 default is to store register values as @code{void *} type.
3696 @defmac ASSUME_EXTENDED_UNWIND_CONTEXT
3698 Define this macro to be 1 if the target always uses extended unwind
3699 context with version, args_size and by_value fields. If it is undefined,
3700 it will be defined to 1 when @code{REG_VALUE_IN_UNWIND_CONTEXT} is
3701 defined and 0 otherwise.
3706 @subsection Eliminating Frame Pointer and Arg Pointer
3708 @c prevent bad page break with this line
3709 This is about eliminating the frame pointer and arg pointer.
3711 @hook TARGET_FRAME_POINTER_REQUIRED
3712 This target hook should return @code{true} if a function must have and use
3713 a frame pointer. This target hook is called in the reload pass. If its return
3714 value is @code{true} the function will have a frame pointer.
3716 This target hook can in principle examine the current function and decide
3717 according to the facts, but on most machines the constant @code{false} or the
3718 constant @code{true} suffices. Use @code{false} when the machine allows code
3719 to be generated with no frame pointer, and doing so saves some time or space.
3720 Use @code{true} when there is no possible advantage to avoiding a frame
3723 In certain cases, the compiler does not know how to produce valid code
3724 without a frame pointer. The compiler recognizes those cases and
3725 automatically gives the function a frame pointer regardless of what
3726 @code{TARGET_FRAME_POINTER_REQUIRED} returns. You don't need to worry about
3729 In a function that does not require a frame pointer, the frame pointer
3730 register can be allocated for ordinary usage, unless you mark it as a
3731 fixed register. See @code{FIXED_REGISTERS} for more information.
3733 Default return value is @code{false}.
3736 @findex get_frame_size
3737 @defmac INITIAL_FRAME_POINTER_OFFSET (@var{depth-var})
3738 A C statement to store in the variable @var{depth-var} the difference
3739 between the frame pointer and the stack pointer values immediately after
3740 the function prologue. The value would be computed from information
3741 such as the result of @code{get_frame_size ()} and the tables of
3742 registers @code{regs_ever_live} and @code{call_used_regs}.
3744 If @code{ELIMINABLE_REGS} is defined, this macro will be not be used and
3745 need not be defined. Otherwise, it must be defined even if
3746 @code{TARGET_FRAME_POINTER_REQUIRED} always returns true; in that
3747 case, you may set @var{depth-var} to anything.
3750 @defmac ELIMINABLE_REGS
3751 If defined, this macro specifies a table of register pairs used to
3752 eliminate unneeded registers that point into the stack frame. If it is not
3753 defined, the only elimination attempted by the compiler is to replace
3754 references to the frame pointer with references to the stack pointer.
3756 The definition of this macro is a list of structure initializations, each
3757 of which specifies an original and replacement register.
3759 On some machines, the position of the argument pointer is not known until
3760 the compilation is completed. In such a case, a separate hard register
3761 must be used for the argument pointer. This register can be eliminated by
3762 replacing it with either the frame pointer or the argument pointer,
3763 depending on whether or not the frame pointer has been eliminated.
3765 In this case, you might specify:
3767 #define ELIMINABLE_REGS \
3768 @{@{ARG_POINTER_REGNUM, STACK_POINTER_REGNUM@}, \
3769 @{ARG_POINTER_REGNUM, FRAME_POINTER_REGNUM@}, \
3770 @{FRAME_POINTER_REGNUM, STACK_POINTER_REGNUM@}@}
3773 Note that the elimination of the argument pointer with the stack pointer is
3774 specified first since that is the preferred elimination.
3777 @hook TARGET_CAN_ELIMINATE
3778 This target hook should returns @code{true} if the compiler is allowed to
3779 try to replace register number @var{from_reg} with register number
3780 @var{to_reg}. This target hook need only be defined if @code{ELIMINABLE_REGS}
3781 is defined, and will usually be @code{true}, since most of the cases
3782 preventing register elimination are things that the compiler already
3785 Default return value is @code{true}.
3788 @defmac INITIAL_ELIMINATION_OFFSET (@var{from-reg}, @var{to-reg}, @var{offset-var})
3789 This macro is similar to @code{INITIAL_FRAME_POINTER_OFFSET}. It
3790 specifies the initial difference between the specified pair of
3791 registers. This macro must be defined if @code{ELIMINABLE_REGS} is
3795 @node Stack Arguments
3796 @subsection Passing Function Arguments on the Stack
3797 @cindex arguments on stack
3798 @cindex stack arguments
3800 The macros in this section control how arguments are passed
3801 on the stack. See the following section for other macros that
3802 control passing certain arguments in registers.
3804 @hook TARGET_PROMOTE_PROTOTYPES
3805 This target hook returns @code{true} if an argument declared in a
3806 prototype as an integral type smaller than @code{int} should actually be
3807 passed as an @code{int}. In addition to avoiding errors in certain
3808 cases of mismatch, it also makes for better code on certain machines.
3809 The default is to not promote prototypes.
3813 A C expression. If nonzero, push insns will be used to pass
3815 If the target machine does not have a push instruction, set it to zero.
3816 That directs GCC to use an alternate strategy: to
3817 allocate the entire argument block and then store the arguments into
3818 it. When @code{PUSH_ARGS} is nonzero, @code{PUSH_ROUNDING} must be defined too.
3821 @defmac PUSH_ARGS_REVERSED
3822 A C expression. If nonzero, function arguments will be evaluated from
3823 last to first, rather than from first to last. If this macro is not
3824 defined, it defaults to @code{PUSH_ARGS} on targets where the stack
3825 and args grow in opposite directions, and 0 otherwise.
3828 @defmac PUSH_ROUNDING (@var{npushed})
3829 A C expression that is the number of bytes actually pushed onto the
3830 stack when an instruction attempts to push @var{npushed} bytes.
3832 On some machines, the definition
3835 #define PUSH_ROUNDING(BYTES) (BYTES)
3839 will suffice. But on other machines, instructions that appear
3840 to push one byte actually push two bytes in an attempt to maintain
3841 alignment. Then the definition should be
3844 #define PUSH_ROUNDING(BYTES) (((BYTES) + 1) & ~1)
3847 If the value of this macro has a type, it should be an unsigned type.
3850 @findex current_function_outgoing_args_size
3851 @defmac ACCUMULATE_OUTGOING_ARGS
3852 A C expression. If nonzero, the maximum amount of space required for outgoing arguments
3853 will be computed and placed into the variable
3854 @code{current_function_outgoing_args_size}. No space will be pushed
3855 onto the stack for each call; instead, the function prologue should
3856 increase the stack frame size by this amount.
3858 Setting both @code{PUSH_ARGS} and @code{ACCUMULATE_OUTGOING_ARGS}
3862 @defmac REG_PARM_STACK_SPACE (@var{fndecl})
3863 Define this macro if functions should assume that stack space has been
3864 allocated for arguments even when their values are passed in
3867 The value of this macro is the size, in bytes, of the area reserved for
3868 arguments passed in registers for the function represented by @var{fndecl},
3869 which can be zero if GCC is calling a library function.
3870 The argument @var{fndecl} can be the FUNCTION_DECL, or the type itself
3873 This space can be allocated by the caller, or be a part of the
3874 machine-dependent stack frame: @code{OUTGOING_REG_PARM_STACK_SPACE} says
3877 @c above is overfull. not sure what to do. --mew 5feb93 did
3878 @c something, not sure if it looks good. --mew 10feb93
3880 @defmac OUTGOING_REG_PARM_STACK_SPACE (@var{fntype})
3881 Define this to a nonzero value if it is the responsibility of the
3882 caller to allocate the area reserved for arguments passed in registers
3883 when calling a function of @var{fntype}. @var{fntype} may be NULL
3884 if the function called is a library function.
3886 If @code{ACCUMULATE_OUTGOING_ARGS} is defined, this macro controls
3887 whether the space for these arguments counts in the value of
3888 @code{current_function_outgoing_args_size}.
3891 @defmac STACK_PARMS_IN_REG_PARM_AREA
3892 Define this macro if @code{REG_PARM_STACK_SPACE} is defined, but the
3893 stack parameters don't skip the area specified by it.
3894 @c i changed this, makes more sens and it should have taken care of the
3895 @c overfull.. not as specific, tho. --mew 5feb93
3897 Normally, when a parameter is not passed in registers, it is placed on the
3898 stack beyond the @code{REG_PARM_STACK_SPACE} area. Defining this macro
3899 suppresses this behavior and causes the parameter to be passed on the
3900 stack in its natural location.
3903 @hook TARGET_RETURN_POPS_ARGS
3904 This target hook returns the number of bytes of its own arguments that
3905 a function pops on returning, or 0 if the function pops no arguments
3906 and the caller must therefore pop them all after the function returns.
3908 @var{fundecl} is a C variable whose value is a tree node that describes
3909 the function in question. Normally it is a node of type
3910 @code{FUNCTION_DECL} that describes the declaration of the function.
3911 From this you can obtain the @code{DECL_ATTRIBUTES} of the function.
3913 @var{funtype} is a C variable whose value is a tree node that
3914 describes the function in question. Normally it is a node of type
3915 @code{FUNCTION_TYPE} that describes the data type of the function.
3916 From this it is possible to obtain the data types of the value and
3917 arguments (if known).
3919 When a call to a library function is being considered, @var{fundecl}
3920 will contain an identifier node for the library function. Thus, if
3921 you need to distinguish among various library functions, you can do so
3922 by their names. Note that ``library function'' in this context means
3923 a function used to perform arithmetic, whose name is known specially
3924 in the compiler and was not mentioned in the C code being compiled.
3926 @var{size} is the number of bytes of arguments passed on the
3927 stack. If a variable number of bytes is passed, it is zero, and
3928 argument popping will always be the responsibility of the calling function.
3930 On the VAX, all functions always pop their arguments, so the definition
3931 of this macro is @var{size}. On the 68000, using the standard
3932 calling convention, no functions pop their arguments, so the value of
3933 the macro is always 0 in this case. But an alternative calling
3934 convention is available in which functions that take a fixed number of
3935 arguments pop them but other functions (such as @code{printf}) pop
3936 nothing (the caller pops all). When this convention is in use,
3937 @var{funtype} is examined to determine whether a function takes a fixed
3938 number of arguments.
3941 @defmac CALL_POPS_ARGS (@var{cum})
3942 A C expression that should indicate the number of bytes a call sequence
3943 pops off the stack. It is added to the value of @code{RETURN_POPS_ARGS}
3944 when compiling a function call.
3946 @var{cum} is the variable in which all arguments to the called function
3947 have been accumulated.
3949 On certain architectures, such as the SH5, a call trampoline is used
3950 that pops certain registers off the stack, depending on the arguments
3951 that have been passed to the function. Since this is a property of the
3952 call site, not of the called function, @code{RETURN_POPS_ARGS} is not
3956 @node Register Arguments
3957 @subsection Passing Arguments in Registers
3958 @cindex arguments in registers
3959 @cindex registers arguments
3961 This section describes the macros which let you control how various
3962 types of arguments are passed in registers or how they are arranged in
3965 @hook TARGET_FUNCTION_ARG
3966 Return an RTX indicating whether a function argument is passed in a
3967 register and if so, which register.
3969 The arguments are @var{ca}, which summarizes all the previous
3970 arguments; @var{mode}, the machine mode of the argument; @var{type},
3971 the data type of the argument as a tree node or 0 if that is not known
3972 (which happens for C support library functions); and @var{named},
3973 which is @code{true} for an ordinary argument and @code{false} for
3974 nameless arguments that correspond to @samp{@dots{}} in the called
3975 function's prototype. @var{type} can be an incomplete type if a
3976 syntax error has previously occurred.
3978 The return value is usually either a @code{reg} RTX for the hard
3979 register in which to pass the argument, or zero to pass the argument
3982 The value of the expression can also be a @code{parallel} RTX@. This is
3983 used when an argument is passed in multiple locations. The mode of the
3984 @code{parallel} should be the mode of the entire argument. The
3985 @code{parallel} holds any number of @code{expr_list} pairs; each one
3986 describes where part of the argument is passed. In each
3987 @code{expr_list} the first operand must be a @code{reg} RTX for the hard
3988 register in which to pass this part of the argument, and the mode of the
3989 register RTX indicates how large this part of the argument is. The
3990 second operand of the @code{expr_list} is a @code{const_int} which gives
3991 the offset in bytes into the entire argument of where this part starts.
3992 As a special exception the first @code{expr_list} in the @code{parallel}
3993 RTX may have a first operand of zero. This indicates that the entire
3994 argument is also stored on the stack.
3996 The last time this hook is called, it is called with @code{MODE ==
3997 VOIDmode}, and its result is passed to the @code{call} or @code{call_value}
3998 pattern as operands 2 and 3 respectively.
4000 @cindex @file{stdarg.h} and register arguments
4001 The usual way to make the ISO library @file{stdarg.h} work on a
4002 machine where some arguments are usually passed in registers, is to
4003 cause nameless arguments to be passed on the stack instead. This is
4004 done by making @code{TARGET_FUNCTION_ARG} return 0 whenever
4005 @var{named} is @code{false}.
4007 @cindex @code{TARGET_MUST_PASS_IN_STACK}, and @code{TARGET_FUNCTION_ARG}
4008 @cindex @code{REG_PARM_STACK_SPACE}, and @code{TARGET_FUNCTION_ARG}
4009 You may use the hook @code{targetm.calls.must_pass_in_stack}
4010 in the definition of this macro to determine if this argument is of a
4011 type that must be passed in the stack. If @code{REG_PARM_STACK_SPACE}
4012 is not defined and @code{TARGET_FUNCTION_ARG} returns nonzero for such an
4013 argument, the compiler will abort. If @code{REG_PARM_STACK_SPACE} is
4014 defined, the argument will be computed in the stack and then loaded into
4018 @hook TARGET_MUST_PASS_IN_STACK
4019 This target hook should return @code{true} if we should not pass @var{type}
4020 solely in registers. The file @file{expr.h} defines a
4021 definition that is usually appropriate, refer to @file{expr.h} for additional
4025 @hook TARGET_FUNCTION_INCOMING_ARG
4026 Define this hook if the target machine has ``register windows'', so
4027 that the register in which a function sees an arguments is not
4028 necessarily the same as the one in which the caller passed the
4031 For such machines, @code{TARGET_FUNCTION_ARG} computes the register in
4032 which the caller passes the value, and
4033 @code{TARGET_FUNCTION_INCOMING_ARG} should be defined in a similar
4034 fashion to tell the function being called where the arguments will
4037 If @code{TARGET_FUNCTION_INCOMING_ARG} is not defined,
4038 @code{TARGET_FUNCTION_ARG} serves both purposes.
4041 @hook TARGET_ARG_PARTIAL_BYTES
4042 This target hook returns the number of bytes at the beginning of an
4043 argument that must be put in registers. The value must be zero for
4044 arguments that are passed entirely in registers or that are entirely
4045 pushed on the stack.
4047 On some machines, certain arguments must be passed partially in
4048 registers and partially in memory. On these machines, typically the
4049 first few words of arguments are passed in registers, and the rest
4050 on the stack. If a multi-word argument (a @code{double} or a
4051 structure) crosses that boundary, its first few words must be passed
4052 in registers and the rest must be pushed. This macro tells the
4053 compiler when this occurs, and how many bytes should go in registers.
4055 @code{TARGET_FUNCTION_ARG} for these arguments should return the first
4056 register to be used by the caller for this argument; likewise
4057 @code{TARGET_FUNCTION_INCOMING_ARG}, for the called function.
4060 @hook TARGET_PASS_BY_REFERENCE
4061 This target hook should return @code{true} if an argument at the
4062 position indicated by @var{cum} should be passed by reference. This
4063 predicate is queried after target independent reasons for being
4064 passed by reference, such as @code{TREE_ADDRESSABLE (type)}.
4066 If the hook returns true, a copy of that argument is made in memory and a
4067 pointer to the argument is passed instead of the argument itself.
4068 The pointer is passed in whatever way is appropriate for passing a pointer
4072 @hook TARGET_CALLEE_COPIES
4073 The function argument described by the parameters to this hook is
4074 known to be passed by reference. The hook should return true if the
4075 function argument should be copied by the callee instead of copied
4078 For any argument for which the hook returns true, if it can be
4079 determined that the argument is not modified, then a copy need
4082 The default version of this hook always returns false.
4085 @defmac CUMULATIVE_ARGS
4086 A C type for declaring a variable that is used as the first argument
4087 of @code{TARGET_FUNCTION_ARG} and other related values. For some
4088 target machines, the type @code{int} suffices and can hold the number
4089 of bytes of argument so far.
4091 There is no need to record in @code{CUMULATIVE_ARGS} anything about the
4092 arguments that have been passed on the stack. The compiler has other
4093 variables to keep track of that. For target machines on which all
4094 arguments are passed on the stack, there is no need to store anything in
4095 @code{CUMULATIVE_ARGS}; however, the data structure must exist and
4096 should not be empty, so use @code{int}.
4099 @defmac OVERRIDE_ABI_FORMAT (@var{fndecl})
4100 If defined, this macro is called before generating any code for a
4101 function, but after the @var{cfun} descriptor for the function has been
4102 created. The back end may use this macro to update @var{cfun} to
4103 reflect an ABI other than that which would normally be used by default.
4104 If the compiler is generating code for a compiler-generated function,
4105 @var{fndecl} may be @code{NULL}.
4108 @defmac INIT_CUMULATIVE_ARGS (@var{cum}, @var{fntype}, @var{libname}, @var{fndecl}, @var{n_named_args})
4109 A C statement (sans semicolon) for initializing the variable
4110 @var{cum} for the state at the beginning of the argument list. The
4111 variable has type @code{CUMULATIVE_ARGS}. The value of @var{fntype}
4112 is the tree node for the data type of the function which will receive
4113 the args, or 0 if the args are to a compiler support library function.
4114 For direct calls that are not libcalls, @var{fndecl} contain the
4115 declaration node of the function. @var{fndecl} is also set when
4116 @code{INIT_CUMULATIVE_ARGS} is used to find arguments for the function
4117 being compiled. @var{n_named_args} is set to the number of named
4118 arguments, including a structure return address if it is passed as a
4119 parameter, when making a call. When processing incoming arguments,
4120 @var{n_named_args} is set to @minus{}1.
4122 When processing a call to a compiler support library function,
4123 @var{libname} identifies which one. It is a @code{symbol_ref} rtx which
4124 contains the name of the function, as a string. @var{libname} is 0 when
4125 an ordinary C function call is being processed. Thus, each time this
4126 macro is called, either @var{libname} or @var{fntype} is nonzero, but
4127 never both of them at once.
4130 @defmac INIT_CUMULATIVE_LIBCALL_ARGS (@var{cum}, @var{mode}, @var{libname})
4131 Like @code{INIT_CUMULATIVE_ARGS} but only used for outgoing libcalls,
4132 it gets a @code{MODE} argument instead of @var{fntype}, that would be
4133 @code{NULL}. @var{indirect} would always be zero, too. If this macro
4134 is not defined, @code{INIT_CUMULATIVE_ARGS (cum, NULL_RTX, libname,
4135 0)} is used instead.
4138 @defmac INIT_CUMULATIVE_INCOMING_ARGS (@var{cum}, @var{fntype}, @var{libname})
4139 Like @code{INIT_CUMULATIVE_ARGS} but overrides it for the purposes of
4140 finding the arguments for the function being compiled. If this macro is
4141 undefined, @code{INIT_CUMULATIVE_ARGS} is used instead.
4143 The value passed for @var{libname} is always 0, since library routines
4144 with special calling conventions are never compiled with GCC@. The
4145 argument @var{libname} exists for symmetry with
4146 @code{INIT_CUMULATIVE_ARGS}.
4147 @c could use "this macro" in place of @code{INIT_CUMULATIVE_ARGS}, maybe.
4148 @c --mew 5feb93 i switched the order of the sentences. --mew 10feb93
4151 @hook TARGET_FUNCTION_ARG_ADVANCE
4152 This hook updates the summarizer variable pointed to by @var{ca} to
4153 advance past an argument in the argument list. The values @var{mode},
4154 @var{type} and @var{named} describe that argument. Once this is done,
4155 the variable @var{cum} is suitable for analyzing the @emph{following}
4156 argument with @code{TARGET_FUNCTION_ARG}, etc.
4158 This hook need not do anything if the argument in question was passed
4159 on the stack. The compiler knows how to track the amount of stack space
4160 used for arguments without any special help.
4163 @defmac FUNCTION_ARG_OFFSET (@var{mode}, @var{type})
4164 If defined, a C expression that is the number of bytes to add to the
4165 offset of the argument passed in memory. This is needed for the SPU,
4166 which passes @code{char} and @code{short} arguments in the preferred
4167 slot that is in the middle of the quad word instead of starting at the
4171 @defmac FUNCTION_ARG_PADDING (@var{mode}, @var{type})
4172 If defined, a C expression which determines whether, and in which direction,
4173 to pad out an argument with extra space. The value should be of type
4174 @code{enum direction}: either @code{upward} to pad above the argument,
4175 @code{downward} to pad below, or @code{none} to inhibit padding.
4177 The @emph{amount} of padding is not controlled by this macro, but by the
4178 target hook @code{TARGET_FUNCTION_ARG_ROUND_BOUNDARY}. It is
4179 always just enough to reach the next multiple of that boundary.
4181 This macro has a default definition which is right for most systems.
4182 For little-endian machines, the default is to pad upward. For
4183 big-endian machines, the default is to pad downward for an argument of
4184 constant size shorter than an @code{int}, and upward otherwise.
4187 @defmac PAD_VARARGS_DOWN
4188 If defined, a C expression which determines whether the default
4189 implementation of va_arg will attempt to pad down before reading the
4190 next argument, if that argument is smaller than its aligned space as
4191 controlled by @code{PARM_BOUNDARY}. If this macro is not defined, all such
4192 arguments are padded down if @code{BYTES_BIG_ENDIAN} is true.
4195 @defmac BLOCK_REG_PADDING (@var{mode}, @var{type}, @var{first})
4196 Specify padding for the last element of a block move between registers and
4197 memory. @var{first} is nonzero if this is the only element. Defining this
4198 macro allows better control of register function parameters on big-endian
4199 machines, without using @code{PARALLEL} rtl. In particular,
4200 @code{MUST_PASS_IN_STACK} need not test padding and mode of types in
4201 registers, as there is no longer a "wrong" part of a register; For example,
4202 a three byte aggregate may be passed in the high part of a register if so
4206 @hook TARGET_FUNCTION_ARG_BOUNDARY
4207 This hook returns the alignment boundary, in bits, of an argument
4208 with the specified mode and type. The default hook returns
4209 @code{PARM_BOUNDARY} for all arguments.
4212 @hook TARGET_FUNCTION_ARG_ROUND_BOUNDARY
4214 @defmac FUNCTION_ARG_REGNO_P (@var{regno})
4215 A C expression that is nonzero if @var{regno} is the number of a hard
4216 register in which function arguments are sometimes passed. This does
4217 @emph{not} include implicit arguments such as the static chain and
4218 the structure-value address. On many machines, no registers can be
4219 used for this purpose since all function arguments are pushed on the
4223 @hook TARGET_SPLIT_COMPLEX_ARG
4224 This hook should return true if parameter of type @var{type} are passed
4225 as two scalar parameters. By default, GCC will attempt to pack complex
4226 arguments into the target's word size. Some ABIs require complex arguments
4227 to be split and treated as their individual components. For example, on
4228 AIX64, complex floats should be passed in a pair of floating point
4229 registers, even though a complex float would fit in one 64-bit floating
4232 The default value of this hook is @code{NULL}, which is treated as always
4236 @hook TARGET_BUILD_BUILTIN_VA_LIST
4237 This hook returns a type node for @code{va_list} for the target.
4238 The default version of the hook returns @code{void*}.
4241 @hook TARGET_ENUM_VA_LIST_P
4242 This target hook is used in function @code{c_common_nodes_and_builtins}
4243 to iterate through the target specific builtin types for va_list. The
4244 variable @var{idx} is used as iterator. @var{pname} has to be a pointer
4245 to a @code{const char *} and @var{ptree} a pointer to a @code{tree} typed
4247 The arguments @var{pname} and @var{ptree} are used to store the result of
4248 this macro and are set to the name of the va_list builtin type and its
4250 If the return value of this macro is zero, then there is no more element.
4251 Otherwise the @var{IDX} should be increased for the next call of this
4252 macro to iterate through all types.
4255 @hook TARGET_FN_ABI_VA_LIST
4256 This hook returns the va_list type of the calling convention specified by
4258 The default version of this hook returns @code{va_list_type_node}.
4261 @hook TARGET_CANONICAL_VA_LIST_TYPE
4262 This hook returns the va_list type of the calling convention specified by the
4263 type of @var{type}. If @var{type} is not a valid va_list type, it returns
4267 @hook TARGET_GIMPLIFY_VA_ARG_EXPR
4268 This hook performs target-specific gimplification of
4269 @code{VA_ARG_EXPR}. The first two parameters correspond to the
4270 arguments to @code{va_arg}; the latter two are as in
4271 @code{gimplify.c:gimplify_expr}.
4274 @hook TARGET_VALID_POINTER_MODE
4275 Define this to return nonzero if the port can handle pointers
4276 with machine mode @var{mode}. The default version of this
4277 hook returns true for both @code{ptr_mode} and @code{Pmode}.
4280 @hook TARGET_REF_MAY_ALIAS_ERRNO
4282 @hook TARGET_SCALAR_MODE_SUPPORTED_P
4283 Define this to return nonzero if the port is prepared to handle
4284 insns involving scalar mode @var{mode}. For a scalar mode to be
4285 considered supported, all the basic arithmetic and comparisons
4288 The default version of this hook returns true for any mode
4289 required to handle the basic C types (as defined by the port).
4290 Included here are the double-word arithmetic supported by the
4291 code in @file{optabs.c}.
4294 @hook TARGET_VECTOR_MODE_SUPPORTED_P
4295 Define this to return nonzero if the port is prepared to handle
4296 insns involving vector mode @var{mode}. At the very least, it
4297 must have move patterns for this mode.
4300 @hook TARGET_ARRAY_MODE_SUPPORTED_P
4302 @hook TARGET_SMALL_REGISTER_CLASSES_FOR_MODE_P
4303 Define this to return nonzero for machine modes for which the port has
4304 small register classes. If this target hook returns nonzero for a given
4305 @var{mode}, the compiler will try to minimize the lifetime of registers
4306 in @var{mode}. The hook may be called with @code{VOIDmode} as argument.
4307 In this case, the hook is expected to return nonzero if it returns nonzero
4310 On some machines, it is risky to let hard registers live across arbitrary
4311 insns. Typically, these machines have instructions that require values
4312 to be in specific registers (like an accumulator), and reload will fail
4313 if the required hard register is used for another purpose across such an
4316 Passes before reload do not know which hard registers will be used
4317 in an instruction, but the machine modes of the registers set or used in
4318 the instruction are already known. And for some machines, register
4319 classes are small for, say, integer registers but not for floating point
4320 registers. For example, the AMD x86-64 architecture requires specific
4321 registers for the legacy x86 integer instructions, but there are many
4322 SSE registers for floating point operations. On such targets, a good
4323 strategy may be to return nonzero from this hook for @code{INTEGRAL_MODE_P}
4324 machine modes but zero for the SSE register classes.
4326 The default version of this hook returns false for any mode. It is always
4327 safe to redefine this hook to return with a nonzero value. But if you
4328 unnecessarily define it, you will reduce the amount of optimizations
4329 that can be performed in some cases. If you do not define this hook
4330 to return a nonzero value when it is required, the compiler will run out
4331 of spill registers and print a fatal error message.
4334 @hook TARGET_FLAGS_REGNUM
4337 @subsection How Scalar Function Values Are Returned
4338 @cindex return values in registers
4339 @cindex values, returned by functions
4340 @cindex scalars, returned as values
4342 This section discusses the macros that control returning scalars as
4343 values---values that can fit in registers.
4345 @hook TARGET_FUNCTION_VALUE
4347 Define this to return an RTX representing the place where a function
4348 returns or receives a value of data type @var{ret_type}, a tree node
4349 representing a data type. @var{fn_decl_or_type} is a tree node
4350 representing @code{FUNCTION_DECL} or @code{FUNCTION_TYPE} of a
4351 function being called. If @var{outgoing} is false, the hook should
4352 compute the register in which the caller will see the return value.
4353 Otherwise, the hook should return an RTX representing the place where
4354 a function returns a value.
4356 On many machines, only @code{TYPE_MODE (@var{ret_type})} is relevant.
4357 (Actually, on most machines, scalar values are returned in the same
4358 place regardless of mode.) The value of the expression is usually a
4359 @code{reg} RTX for the hard register where the return value is stored.
4360 The value can also be a @code{parallel} RTX, if the return value is in
4361 multiple places. See @code{TARGET_FUNCTION_ARG} for an explanation of the
4362 @code{parallel} form. Note that the callee will populate every
4363 location specified in the @code{parallel}, but if the first element of
4364 the @code{parallel} contains the whole return value, callers will use
4365 that element as the canonical location and ignore the others. The m68k
4366 port uses this type of @code{parallel} to return pointers in both
4367 @samp{%a0} (the canonical location) and @samp{%d0}.
4369 If @code{TARGET_PROMOTE_FUNCTION_RETURN} returns true, you must apply
4370 the same promotion rules specified in @code{PROMOTE_MODE} if
4371 @var{valtype} is a scalar type.
4373 If the precise function being called is known, @var{func} is a tree
4374 node (@code{FUNCTION_DECL}) for it; otherwise, @var{func} is a null
4375 pointer. This makes it possible to use a different value-returning
4376 convention for specific functions when all their calls are
4379 Some target machines have ``register windows'' so that the register in
4380 which a function returns its value is not the same as the one in which
4381 the caller sees the value. For such machines, you should return
4382 different RTX depending on @var{outgoing}.
4384 @code{TARGET_FUNCTION_VALUE} is not used for return values with
4385 aggregate data types, because these are returned in another way. See
4386 @code{TARGET_STRUCT_VALUE_RTX} and related macros, below.
4389 @defmac FUNCTION_VALUE (@var{valtype}, @var{func})
4390 This macro has been deprecated. Use @code{TARGET_FUNCTION_VALUE} for
4391 a new target instead.
4394 @defmac LIBCALL_VALUE (@var{mode})
4395 A C expression to create an RTX representing the place where a library
4396 function returns a value of mode @var{mode}.
4398 Note that ``library function'' in this context means a compiler
4399 support routine, used to perform arithmetic, whose name is known
4400 specially by the compiler and was not mentioned in the C code being
4404 @hook TARGET_LIBCALL_VALUE
4405 Define this hook if the back-end needs to know the name of the libcall
4406 function in order to determine where the result should be returned.
4408 The mode of the result is given by @var{mode} and the name of the called
4409 library function is given by @var{fun}. The hook should return an RTX
4410 representing the place where the library function result will be returned.
4412 If this hook is not defined, then LIBCALL_VALUE will be used.
4415 @defmac FUNCTION_VALUE_REGNO_P (@var{regno})
4416 A C expression that is nonzero if @var{regno} is the number of a hard
4417 register in which the values of called function may come back.
4419 A register whose use for returning values is limited to serving as the
4420 second of a pair (for a value of type @code{double}, say) need not be
4421 recognized by this macro. So for most machines, this definition
4425 #define FUNCTION_VALUE_REGNO_P(N) ((N) == 0)
4428 If the machine has register windows, so that the caller and the called
4429 function use different registers for the return value, this macro
4430 should recognize only the caller's register numbers.
4432 This macro has been deprecated. Use @code{TARGET_FUNCTION_VALUE_REGNO_P}
4433 for a new target instead.
4436 @hook TARGET_FUNCTION_VALUE_REGNO_P
4437 A target hook that return @code{true} if @var{regno} is the number of a hard
4438 register in which the values of called function may come back.
4440 A register whose use for returning values is limited to serving as the
4441 second of a pair (for a value of type @code{double}, say) need not be
4442 recognized by this target hook.
4444 If the machine has register windows, so that the caller and the called
4445 function use different registers for the return value, this target hook
4446 should recognize only the caller's register numbers.
4448 If this hook is not defined, then FUNCTION_VALUE_REGNO_P will be used.
4451 @defmac APPLY_RESULT_SIZE
4452 Define this macro if @samp{untyped_call} and @samp{untyped_return}
4453 need more space than is implied by @code{FUNCTION_VALUE_REGNO_P} for
4454 saving and restoring an arbitrary return value.
4457 @hook TARGET_RETURN_IN_MSB
4458 This hook should return true if values of type @var{type} are returned
4459 at the most significant end of a register (in other words, if they are
4460 padded at the least significant end). You can assume that @var{type}
4461 is returned in a register; the caller is required to check this.
4463 Note that the register provided by @code{TARGET_FUNCTION_VALUE} must
4464 be able to hold the complete return value. For example, if a 1-, 2-
4465 or 3-byte structure is returned at the most significant end of a
4466 4-byte register, @code{TARGET_FUNCTION_VALUE} should provide an
4470 @node Aggregate Return
4471 @subsection How Large Values Are Returned
4472 @cindex aggregates as return values
4473 @cindex large return values
4474 @cindex returning aggregate values
4475 @cindex structure value address
4477 When a function value's mode is @code{BLKmode} (and in some other
4478 cases), the value is not returned according to
4479 @code{TARGET_FUNCTION_VALUE} (@pxref{Scalar Return}). Instead, the
4480 caller passes the address of a block of memory in which the value
4481 should be stored. This address is called the @dfn{structure value
4484 This section describes how to control returning structure values in
4487 @hook TARGET_RETURN_IN_MEMORY
4488 This target hook should return a nonzero value to say to return the
4489 function value in memory, just as large structures are always returned.
4490 Here @var{type} will be the data type of the value, and @var{fntype}
4491 will be the type of the function doing the returning, or @code{NULL} for
4494 Note that values of mode @code{BLKmode} must be explicitly handled
4495 by this function. Also, the option @option{-fpcc-struct-return}
4496 takes effect regardless of this macro. On most systems, it is
4497 possible to leave the hook undefined; this causes a default
4498 definition to be used, whose value is the constant 1 for @code{BLKmode}
4499 values, and 0 otherwise.
4501 Do not use this hook to indicate that structures and unions should always
4502 be returned in memory. You should instead use @code{DEFAULT_PCC_STRUCT_RETURN}
4506 @defmac DEFAULT_PCC_STRUCT_RETURN
4507 Define this macro to be 1 if all structure and union return values must be
4508 in memory. Since this results in slower code, this should be defined
4509 only if needed for compatibility with other compilers or with an ABI@.
4510 If you define this macro to be 0, then the conventions used for structure
4511 and union return values are decided by the @code{TARGET_RETURN_IN_MEMORY}
4514 If not defined, this defaults to the value 1.
4517 @hook TARGET_STRUCT_VALUE_RTX
4518 This target hook should return the location of the structure value
4519 address (normally a @code{mem} or @code{reg}), or 0 if the address is
4520 passed as an ``invisible'' first argument. Note that @var{fndecl} may
4521 be @code{NULL}, for libcalls. You do not need to define this target
4522 hook if the address is always passed as an ``invisible'' first
4525 On some architectures the place where the structure value address
4526 is found by the called function is not the same place that the
4527 caller put it. This can be due to register windows, or it could
4528 be because the function prologue moves it to a different place.
4529 @var{incoming} is @code{1} or @code{2} when the location is needed in
4530 the context of the called function, and @code{0} in the context of
4533 If @var{incoming} is nonzero and the address is to be found on the
4534 stack, return a @code{mem} which refers to the frame pointer. If
4535 @var{incoming} is @code{2}, the result is being used to fetch the
4536 structure value address at the beginning of a function. If you need
4537 to emit adjusting code, you should do it at this point.
4540 @defmac PCC_STATIC_STRUCT_RETURN
4541 Define this macro if the usual system convention on the target machine
4542 for returning structures and unions is for the called function to return
4543 the address of a static variable containing the value.
4545 Do not define this if the usual system convention is for the caller to
4546 pass an address to the subroutine.
4548 This macro has effect in @option{-fpcc-struct-return} mode, but it does
4549 nothing when you use @option{-freg-struct-return} mode.
4552 @hook TARGET_GET_RAW_RESULT_MODE
4554 @hook TARGET_GET_RAW_ARG_MODE
4557 @subsection Caller-Saves Register Allocation
4559 If you enable it, GCC can save registers around function calls. This
4560 makes it possible to use call-clobbered registers to hold variables that
4561 must live across calls.
4563 @defmac CALLER_SAVE_PROFITABLE (@var{refs}, @var{calls})
4564 A C expression to determine whether it is worthwhile to consider placing
4565 a pseudo-register in a call-clobbered hard register and saving and
4566 restoring it around each function call. The expression should be 1 when
4567 this is worth doing, and 0 otherwise.
4569 If you don't define this macro, a default is used which is good on most
4570 machines: @code{4 * @var{calls} < @var{refs}}.
4573 @defmac HARD_REGNO_CALLER_SAVE_MODE (@var{regno}, @var{nregs})
4574 A C expression specifying which mode is required for saving @var{nregs}
4575 of a pseudo-register in call-clobbered hard register @var{regno}. If
4576 @var{regno} is unsuitable for caller save, @code{VOIDmode} should be
4577 returned. For most machines this macro need not be defined since GCC
4578 will select the smallest suitable mode.
4581 @node Function Entry
4582 @subsection Function Entry and Exit
4583 @cindex function entry and exit
4587 This section describes the macros that output function entry
4588 (@dfn{prologue}) and exit (@dfn{epilogue}) code.
4590 @hook TARGET_ASM_FUNCTION_PROLOGUE
4591 If defined, a function that outputs the assembler code for entry to a
4592 function. The prologue is responsible for setting up the stack frame,
4593 initializing the frame pointer register, saving registers that must be
4594 saved, and allocating @var{size} additional bytes of storage for the
4595 local variables. @var{size} is an integer. @var{file} is a stdio
4596 stream to which the assembler code should be output.
4598 The label for the beginning of the function need not be output by this
4599 macro. That has already been done when the macro is run.
4601 @findex regs_ever_live
4602 To determine which registers to save, the macro can refer to the array
4603 @code{regs_ever_live}: element @var{r} is nonzero if hard register
4604 @var{r} is used anywhere within the function. This implies the function
4605 prologue should save register @var{r}, provided it is not one of the
4606 call-used registers. (@code{TARGET_ASM_FUNCTION_EPILOGUE} must likewise use
4607 @code{regs_ever_live}.)
4609 On machines that have ``register windows'', the function entry code does
4610 not save on the stack the registers that are in the windows, even if
4611 they are supposed to be preserved by function calls; instead it takes
4612 appropriate steps to ``push'' the register stack, if any non-call-used
4613 registers are used in the function.
4615 @findex frame_pointer_needed
4616 On machines where functions may or may not have frame-pointers, the
4617 function entry code must vary accordingly; it must set up the frame
4618 pointer if one is wanted, and not otherwise. To determine whether a
4619 frame pointer is in wanted, the macro can refer to the variable
4620 @code{frame_pointer_needed}. The variable's value will be 1 at run
4621 time in a function that needs a frame pointer. @xref{Elimination}.
4623 The function entry code is responsible for allocating any stack space
4624 required for the function. This stack space consists of the regions
4625 listed below. In most cases, these regions are allocated in the
4626 order listed, with the last listed region closest to the top of the
4627 stack (the lowest address if @code{STACK_GROWS_DOWNWARD} is defined, and
4628 the highest address if it is not defined). You can use a different order
4629 for a machine if doing so is more convenient or required for
4630 compatibility reasons. Except in cases where required by standard
4631 or by a debugger, there is no reason why the stack layout used by GCC
4632 need agree with that used by other compilers for a machine.
4635 @hook TARGET_ASM_FUNCTION_END_PROLOGUE
4636 If defined, a function that outputs assembler code at the end of a
4637 prologue. This should be used when the function prologue is being
4638 emitted as RTL, and you have some extra assembler that needs to be
4639 emitted. @xref{prologue instruction pattern}.
4642 @hook TARGET_ASM_FUNCTION_BEGIN_EPILOGUE
4643 If defined, a function that outputs assembler code at the start of an
4644 epilogue. This should be used when the function epilogue is being
4645 emitted as RTL, and you have some extra assembler that needs to be
4646 emitted. @xref{epilogue instruction pattern}.
4649 @hook TARGET_ASM_FUNCTION_EPILOGUE
4650 If defined, a function that outputs the assembler code for exit from a
4651 function. The epilogue is responsible for restoring the saved
4652 registers and stack pointer to their values when the function was
4653 called, and returning control to the caller. This macro takes the
4654 same arguments as the macro @code{TARGET_ASM_FUNCTION_PROLOGUE}, and the
4655 registers to restore are determined from @code{regs_ever_live} and
4656 @code{CALL_USED_REGISTERS} in the same way.
4658 On some machines, there is a single instruction that does all the work
4659 of returning from the function. On these machines, give that
4660 instruction the name @samp{return} and do not define the macro
4661 @code{TARGET_ASM_FUNCTION_EPILOGUE} at all.
4663 Do not define a pattern named @samp{return} if you want the
4664 @code{TARGET_ASM_FUNCTION_EPILOGUE} to be used. If you want the target
4665 switches to control whether return instructions or epilogues are used,
4666 define a @samp{return} pattern with a validity condition that tests the
4667 target switches appropriately. If the @samp{return} pattern's validity
4668 condition is false, epilogues will be used.
4670 On machines where functions may or may not have frame-pointers, the
4671 function exit code must vary accordingly. Sometimes the code for these
4672 two cases is completely different. To determine whether a frame pointer
4673 is wanted, the macro can refer to the variable
4674 @code{frame_pointer_needed}. The variable's value will be 1 when compiling
4675 a function that needs a frame pointer.
4677 Normally, @code{TARGET_ASM_FUNCTION_PROLOGUE} and
4678 @code{TARGET_ASM_FUNCTION_EPILOGUE} must treat leaf functions specially.
4679 The C variable @code{current_function_is_leaf} is nonzero for such a
4680 function. @xref{Leaf Functions}.
4682 On some machines, some functions pop their arguments on exit while
4683 others leave that for the caller to do. For example, the 68020 when
4684 given @option{-mrtd} pops arguments in functions that take a fixed
4685 number of arguments.
4687 @findex current_function_pops_args
4688 Your definition of the macro @code{RETURN_POPS_ARGS} decides which
4689 functions pop their own arguments. @code{TARGET_ASM_FUNCTION_EPILOGUE}
4690 needs to know what was decided. The number of bytes of the current
4691 function's arguments that this function should pop is available in
4692 @code{crtl->args.pops_args}. @xref{Scalar Return}.
4697 @findex current_function_pretend_args_size
4698 A region of @code{current_function_pretend_args_size} bytes of
4699 uninitialized space just underneath the first argument arriving on the
4700 stack. (This may not be at the very start of the allocated stack region
4701 if the calling sequence has pushed anything else since pushing the stack
4702 arguments. But usually, on such machines, nothing else has been pushed
4703 yet, because the function prologue itself does all the pushing.) This
4704 region is used on machines where an argument may be passed partly in
4705 registers and partly in memory, and, in some cases to support the
4706 features in @code{<stdarg.h>}.
4709 An area of memory used to save certain registers used by the function.
4710 The size of this area, which may also include space for such things as
4711 the return address and pointers to previous stack frames, is
4712 machine-specific and usually depends on which registers have been used
4713 in the function. Machines with register windows often do not require
4717 A region of at least @var{size} bytes, possibly rounded up to an allocation
4718 boundary, to contain the local variables of the function. On some machines,
4719 this region and the save area may occur in the opposite order, with the
4720 save area closer to the top of the stack.
4723 @cindex @code{ACCUMULATE_OUTGOING_ARGS} and stack frames
4724 Optionally, when @code{ACCUMULATE_OUTGOING_ARGS} is defined, a region of
4725 @code{current_function_outgoing_args_size} bytes to be used for outgoing
4726 argument lists of the function. @xref{Stack Arguments}.
4729 @defmac EXIT_IGNORE_STACK
4730 Define this macro as a C expression that is nonzero if the return
4731 instruction or the function epilogue ignores the value of the stack
4732 pointer; in other words, if it is safe to delete an instruction to
4733 adjust the stack pointer before a return from the function. The
4736 Note that this macro's value is relevant only for functions for which
4737 frame pointers are maintained. It is never safe to delete a final
4738 stack adjustment in a function that has no frame pointer, and the
4739 compiler knows this regardless of @code{EXIT_IGNORE_STACK}.
4742 @defmac EPILOGUE_USES (@var{regno})
4743 Define this macro as a C expression that is nonzero for registers that are
4744 used by the epilogue or the @samp{return} pattern. The stack and frame
4745 pointer registers are already assumed to be used as needed.
4748 @defmac EH_USES (@var{regno})
4749 Define this macro as a C expression that is nonzero for registers that are
4750 used by the exception handling mechanism, and so should be considered live
4751 on entry to an exception edge.
4754 @defmac DELAY_SLOTS_FOR_EPILOGUE
4755 Define this macro if the function epilogue contains delay slots to which
4756 instructions from the rest of the function can be ``moved''. The
4757 definition should be a C expression whose value is an integer
4758 representing the number of delay slots there.
4761 @defmac ELIGIBLE_FOR_EPILOGUE_DELAY (@var{insn}, @var{n})
4762 A C expression that returns 1 if @var{insn} can be placed in delay
4763 slot number @var{n} of the epilogue.
4765 The argument @var{n} is an integer which identifies the delay slot now
4766 being considered (since different slots may have different rules of
4767 eligibility). It is never negative and is always less than the number
4768 of epilogue delay slots (what @code{DELAY_SLOTS_FOR_EPILOGUE} returns).
4769 If you reject a particular insn for a given delay slot, in principle, it
4770 may be reconsidered for a subsequent delay slot. Also, other insns may
4771 (at least in principle) be considered for the so far unfilled delay
4774 @findex current_function_epilogue_delay_list
4775 @findex final_scan_insn
4776 The insns accepted to fill the epilogue delay slots are put in an RTL
4777 list made with @code{insn_list} objects, stored in the variable
4778 @code{current_function_epilogue_delay_list}. The insn for the first
4779 delay slot comes first in the list. Your definition of the macro
4780 @code{TARGET_ASM_FUNCTION_EPILOGUE} should fill the delay slots by
4781 outputting the insns in this list, usually by calling
4782 @code{final_scan_insn}.
4784 You need not define this macro if you did not define
4785 @code{DELAY_SLOTS_FOR_EPILOGUE}.
4788 @hook TARGET_ASM_OUTPUT_MI_THUNK
4789 A function that outputs the assembler code for a thunk
4790 function, used to implement C++ virtual function calls with multiple
4791 inheritance. The thunk acts as a wrapper around a virtual function,
4792 adjusting the implicit object parameter before handing control off to
4795 First, emit code to add the integer @var{delta} to the location that
4796 contains the incoming first argument. Assume that this argument
4797 contains a pointer, and is the one used to pass the @code{this} pointer
4798 in C++. This is the incoming argument @emph{before} the function prologue,
4799 e.g.@: @samp{%o0} on a sparc. The addition must preserve the values of
4800 all other incoming arguments.
4802 Then, if @var{vcall_offset} is nonzero, an additional adjustment should be
4803 made after adding @code{delta}. In particular, if @var{p} is the
4804 adjusted pointer, the following adjustment should be made:
4807 p += (*((ptrdiff_t **)p))[vcall_offset/sizeof(ptrdiff_t)]
4810 After the additions, emit code to jump to @var{function}, which is a
4811 @code{FUNCTION_DECL}. This is a direct pure jump, not a call, and does
4812 not touch the return address. Hence returning from @var{FUNCTION} will
4813 return to whoever called the current @samp{thunk}.
4815 The effect must be as if @var{function} had been called directly with
4816 the adjusted first argument. This macro is responsible for emitting all
4817 of the code for a thunk function; @code{TARGET_ASM_FUNCTION_PROLOGUE}
4818 and @code{TARGET_ASM_FUNCTION_EPILOGUE} are not invoked.
4820 The @var{thunk_fndecl} is redundant. (@var{delta} and @var{function}
4821 have already been extracted from it.) It might possibly be useful on
4822 some targets, but probably not.
4824 If you do not define this macro, the target-independent code in the C++
4825 front end will generate a less efficient heavyweight thunk that calls
4826 @var{function} instead of jumping to it. The generic approach does
4827 not support varargs.
4830 @hook TARGET_ASM_CAN_OUTPUT_MI_THUNK
4831 A function that returns true if TARGET_ASM_OUTPUT_MI_THUNK would be able
4832 to output the assembler code for the thunk function specified by the
4833 arguments it is passed, and false otherwise. In the latter case, the
4834 generic approach will be used by the C++ front end, with the limitations
4839 @subsection Generating Code for Profiling
4840 @cindex profiling, code generation
4842 These macros will help you generate code for profiling.
4844 @defmac FUNCTION_PROFILER (@var{file}, @var{labelno})
4845 A C statement or compound statement to output to @var{file} some
4846 assembler code to call the profiling subroutine @code{mcount}.
4849 The details of how @code{mcount} expects to be called are determined by
4850 your operating system environment, not by GCC@. To figure them out,
4851 compile a small program for profiling using the system's installed C
4852 compiler and look at the assembler code that results.
4854 Older implementations of @code{mcount} expect the address of a counter
4855 variable to be loaded into some register. The name of this variable is
4856 @samp{LP} followed by the number @var{labelno}, so you would generate
4857 the name using @samp{LP%d} in a @code{fprintf}.
4860 @defmac PROFILE_HOOK
4861 A C statement or compound statement to output to @var{file} some assembly
4862 code to call the profiling subroutine @code{mcount} even the target does
4863 not support profiling.
4866 @defmac NO_PROFILE_COUNTERS
4867 Define this macro to be an expression with a nonzero value if the
4868 @code{mcount} subroutine on your system does not need a counter variable
4869 allocated for each function. This is true for almost all modern
4870 implementations. If you define this macro, you must not use the
4871 @var{labelno} argument to @code{FUNCTION_PROFILER}.
4874 @defmac PROFILE_BEFORE_PROLOGUE
4875 Define this macro if the code for function profiling should come before
4876 the function prologue. Normally, the profiling code comes after.
4880 @subsection Permitting tail calls
4883 @hook TARGET_FUNCTION_OK_FOR_SIBCALL
4884 True if it is ok to do sibling call optimization for the specified
4885 call expression @var{exp}. @var{decl} will be the called function,
4886 or @code{NULL} if this is an indirect call.
4888 It is not uncommon for limitations of calling conventions to prevent
4889 tail calls to functions outside the current unit of translation, or
4890 during PIC compilation. The hook is used to enforce these restrictions,
4891 as the @code{sibcall} md pattern can not fail, or fall over to a
4892 ``normal'' call. The criteria for successful sibling call optimization
4893 may vary greatly between different architectures.
4896 @hook TARGET_EXTRA_LIVE_ON_ENTRY
4897 Add any hard registers to @var{regs} that are live on entry to the
4898 function. This hook only needs to be defined to provide registers that
4899 cannot be found by examination of FUNCTION_ARG_REGNO_P, the callee saved
4900 registers, STATIC_CHAIN_INCOMING_REGNUM, STATIC_CHAIN_REGNUM,
4901 TARGET_STRUCT_VALUE_RTX, FRAME_POINTER_REGNUM, EH_USES,
4902 FRAME_POINTER_REGNUM, ARG_POINTER_REGNUM, and the PIC_OFFSET_TABLE_REGNUM.
4905 @node Stack Smashing Protection
4906 @subsection Stack smashing protection
4907 @cindex stack smashing protection
4909 @hook TARGET_STACK_PROTECT_GUARD
4910 This hook returns a @code{DECL} node for the external variable to use
4911 for the stack protection guard. This variable is initialized by the
4912 runtime to some random value and is used to initialize the guard value
4913 that is placed at the top of the local stack frame. The type of this
4914 variable must be @code{ptr_type_node}.
4916 The default version of this hook creates a variable called
4917 @samp{__stack_chk_guard}, which is normally defined in @file{libgcc2.c}.
4920 @hook TARGET_STACK_PROTECT_FAIL
4921 This hook returns a tree expression that alerts the runtime that the
4922 stack protect guard variable has been modified. This expression should
4923 involve a call to a @code{noreturn} function.
4925 The default version of this hook invokes a function called
4926 @samp{__stack_chk_fail}, taking no arguments. This function is
4927 normally defined in @file{libgcc2.c}.
4930 @hook TARGET_SUPPORTS_SPLIT_STACK
4933 @section Implementing the Varargs Macros
4934 @cindex varargs implementation
4936 GCC comes with an implementation of @code{<varargs.h>} and
4937 @code{<stdarg.h>} that work without change on machines that pass arguments
4938 on the stack. Other machines require their own implementations of
4939 varargs, and the two machine independent header files must have
4940 conditionals to include it.
4942 ISO @code{<stdarg.h>} differs from traditional @code{<varargs.h>} mainly in
4943 the calling convention for @code{va_start}. The traditional
4944 implementation takes just one argument, which is the variable in which
4945 to store the argument pointer. The ISO implementation of
4946 @code{va_start} takes an additional second argument. The user is
4947 supposed to write the last named argument of the function here.
4949 However, @code{va_start} should not use this argument. The way to find
4950 the end of the named arguments is with the built-in functions described
4953 @defmac __builtin_saveregs ()
4954 Use this built-in function to save the argument registers in memory so
4955 that the varargs mechanism can access them. Both ISO and traditional
4956 versions of @code{va_start} must use @code{__builtin_saveregs}, unless
4957 you use @code{TARGET_SETUP_INCOMING_VARARGS} (see below) instead.
4959 On some machines, @code{__builtin_saveregs} is open-coded under the
4960 control of the target hook @code{TARGET_EXPAND_BUILTIN_SAVEREGS}. On
4961 other machines, it calls a routine written in assembler language,
4962 found in @file{libgcc2.c}.
4964 Code generated for the call to @code{__builtin_saveregs} appears at the
4965 beginning of the function, as opposed to where the call to
4966 @code{__builtin_saveregs} is written, regardless of what the code is.
4967 This is because the registers must be saved before the function starts
4968 to use them for its own purposes.
4969 @c i rewrote the first sentence above to fix an overfull hbox. --mew
4973 @defmac __builtin_next_arg (@var{lastarg})
4974 This builtin returns the address of the first anonymous stack
4975 argument, as type @code{void *}. If @code{ARGS_GROW_DOWNWARD}, it
4976 returns the address of the location above the first anonymous stack
4977 argument. Use it in @code{va_start} to initialize the pointer for
4978 fetching arguments from the stack. Also use it in @code{va_start} to
4979 verify that the second parameter @var{lastarg} is the last named argument
4980 of the current function.
4983 @defmac __builtin_classify_type (@var{object})
4984 Since each machine has its own conventions for which data types are
4985 passed in which kind of register, your implementation of @code{va_arg}
4986 has to embody these conventions. The easiest way to categorize the
4987 specified data type is to use @code{__builtin_classify_type} together
4988 with @code{sizeof} and @code{__alignof__}.
4990 @code{__builtin_classify_type} ignores the value of @var{object},
4991 considering only its data type. It returns an integer describing what
4992 kind of type that is---integer, floating, pointer, structure, and so on.
4994 The file @file{typeclass.h} defines an enumeration that you can use to
4995 interpret the values of @code{__builtin_classify_type}.
4998 These machine description macros help implement varargs:
5000 @hook TARGET_EXPAND_BUILTIN_SAVEREGS
5001 If defined, this hook produces the machine-specific code for a call to
5002 @code{__builtin_saveregs}. This code will be moved to the very
5003 beginning of the function, before any parameter access are made. The
5004 return value of this function should be an RTX that contains the value
5005 to use as the return of @code{__builtin_saveregs}.
5008 @hook TARGET_SETUP_INCOMING_VARARGS
5009 This target hook offers an alternative to using
5010 @code{__builtin_saveregs} and defining the hook
5011 @code{TARGET_EXPAND_BUILTIN_SAVEREGS}. Use it to store the anonymous
5012 register arguments into the stack so that all the arguments appear to
5013 have been passed consecutively on the stack. Once this is done, you can
5014 use the standard implementation of varargs that works for machines that
5015 pass all their arguments on the stack.
5017 The argument @var{args_so_far} points to the @code{CUMULATIVE_ARGS} data
5018 structure, containing the values that are obtained after processing the
5019 named arguments. The arguments @var{mode} and @var{type} describe the
5020 last named argument---its machine mode and its data type as a tree node.
5022 The target hook should do two things: first, push onto the stack all the
5023 argument registers @emph{not} used for the named arguments, and second,
5024 store the size of the data thus pushed into the @code{int}-valued
5025 variable pointed to by @var{pretend_args_size}. The value that you
5026 store here will serve as additional offset for setting up the stack
5029 Because you must generate code to push the anonymous arguments at
5030 compile time without knowing their data types,
5031 @code{TARGET_SETUP_INCOMING_VARARGS} is only useful on machines that
5032 have just a single category of argument register and use it uniformly
5035 If the argument @var{second_time} is nonzero, it means that the
5036 arguments of the function are being analyzed for the second time. This
5037 happens for an inline function, which is not actually compiled until the
5038 end of the source file. The hook @code{TARGET_SETUP_INCOMING_VARARGS} should
5039 not generate any instructions in this case.
5042 @hook TARGET_STRICT_ARGUMENT_NAMING
5043 Define this hook to return @code{true} if the location where a function
5044 argument is passed depends on whether or not it is a named argument.
5046 This hook controls how the @var{named} argument to @code{TARGET_FUNCTION_ARG}
5047 is set for varargs and stdarg functions. If this hook returns
5048 @code{true}, the @var{named} argument is always true for named
5049 arguments, and false for unnamed arguments. If it returns @code{false},
5050 but @code{TARGET_PRETEND_OUTGOING_VARARGS_NAMED} returns @code{true},
5051 then all arguments are treated as named. Otherwise, all named arguments
5052 except the last are treated as named.
5054 You need not define this hook if it always returns @code{false}.
5057 @hook TARGET_PRETEND_OUTGOING_VARARGS_NAMED
5058 If you need to conditionally change ABIs so that one works with
5059 @code{TARGET_SETUP_INCOMING_VARARGS}, but the other works like neither
5060 @code{TARGET_SETUP_INCOMING_VARARGS} nor @code{TARGET_STRICT_ARGUMENT_NAMING} was
5061 defined, then define this hook to return @code{true} if
5062 @code{TARGET_SETUP_INCOMING_VARARGS} is used, @code{false} otherwise.
5063 Otherwise, you should not define this hook.
5067 @section Trampolines for Nested Functions
5068 @cindex trampolines for nested functions
5069 @cindex nested functions, trampolines for
5071 A @dfn{trampoline} is a small piece of code that is created at run time
5072 when the address of a nested function is taken. It normally resides on
5073 the stack, in the stack frame of the containing function. These macros
5074 tell GCC how to generate code to allocate and initialize a
5077 The instructions in the trampoline must do two things: load a constant
5078 address into the static chain register, and jump to the real address of
5079 the nested function. On CISC machines such as the m68k, this requires
5080 two instructions, a move immediate and a jump. Then the two addresses
5081 exist in the trampoline as word-long immediate operands. On RISC
5082 machines, it is often necessary to load each address into a register in
5083 two parts. Then pieces of each address form separate immediate
5086 The code generated to initialize the trampoline must store the variable
5087 parts---the static chain value and the function address---into the
5088 immediate operands of the instructions. On a CISC machine, this is
5089 simply a matter of copying each address to a memory reference at the
5090 proper offset from the start of the trampoline. On a RISC machine, it
5091 may be necessary to take out pieces of the address and store them
5094 @hook TARGET_ASM_TRAMPOLINE_TEMPLATE
5095 This hook is called by @code{assemble_trampoline_template} to output,
5096 on the stream @var{f}, assembler code for a block of data that contains
5097 the constant parts of a trampoline. This code should not include a
5098 label---the label is taken care of automatically.
5100 If you do not define this hook, it means no template is needed
5101 for the target. Do not define this hook on systems where the block move
5102 code to copy the trampoline into place would be larger than the code
5103 to generate it on the spot.
5106 @defmac TRAMPOLINE_SECTION
5107 Return the section into which the trampoline template is to be placed
5108 (@pxref{Sections}). The default value is @code{readonly_data_section}.
5111 @defmac TRAMPOLINE_SIZE
5112 A C expression for the size in bytes of the trampoline, as an integer.
5115 @defmac TRAMPOLINE_ALIGNMENT
5116 Alignment required for trampolines, in bits.
5118 If you don't define this macro, the value of @code{FUNCTION_ALIGNMENT}
5119 is used for aligning trampolines.
5122 @hook TARGET_TRAMPOLINE_INIT
5123 This hook is called to initialize a trampoline.
5124 @var{m_tramp} is an RTX for the memory block for the trampoline; @var{fndecl}
5125 is the @code{FUNCTION_DECL} for the nested function; @var{static_chain} is an
5126 RTX for the static chain value that should be passed to the function
5129 If the target defines @code{TARGET_ASM_TRAMPOLINE_TEMPLATE}, then the
5130 first thing this hook should do is emit a block move into @var{m_tramp}
5131 from the memory block returned by @code{assemble_trampoline_template}.
5132 Note that the block move need only cover the constant parts of the
5133 trampoline. If the target isolates the variable parts of the trampoline
5134 to the end, not all @code{TRAMPOLINE_SIZE} bytes need be copied.
5136 If the target requires any other actions, such as flushing caches or
5137 enabling stack execution, these actions should be performed after
5138 initializing the trampoline proper.
5141 @hook TARGET_TRAMPOLINE_ADJUST_ADDRESS
5142 This hook should perform any machine-specific adjustment in
5143 the address of the trampoline. Its argument contains the address of the
5144 memory block that was passed to @code{TARGET_TRAMPOLINE_INIT}. In case
5145 the address to be used for a function call should be different from the
5146 address at which the template was stored, the different address should
5147 be returned; otherwise @var{addr} should be returned unchanged.
5148 If this hook is not defined, @var{addr} will be used for function calls.
5151 Implementing trampolines is difficult on many machines because they have
5152 separate instruction and data caches. Writing into a stack location
5153 fails to clear the memory in the instruction cache, so when the program
5154 jumps to that location, it executes the old contents.
5156 Here are two possible solutions. One is to clear the relevant parts of
5157 the instruction cache whenever a trampoline is set up. The other is to
5158 make all trampolines identical, by having them jump to a standard
5159 subroutine. The former technique makes trampoline execution faster; the
5160 latter makes initialization faster.
5162 To clear the instruction cache when a trampoline is initialized, define
5163 the following macro.
5165 @defmac CLEAR_INSN_CACHE (@var{beg}, @var{end})
5166 If defined, expands to a C expression clearing the @emph{instruction
5167 cache} in the specified interval. The definition of this macro would
5168 typically be a series of @code{asm} statements. Both @var{beg} and
5169 @var{end} are both pointer expressions.
5172 To use a standard subroutine, define the following macro. In addition,
5173 you must make sure that the instructions in a trampoline fill an entire
5174 cache line with identical instructions, or else ensure that the
5175 beginning of the trampoline code is always aligned at the same point in
5176 its cache line. Look in @file{m68k.h} as a guide.
5178 @defmac TRANSFER_FROM_TRAMPOLINE
5179 Define this macro if trampolines need a special subroutine to do their
5180 work. The macro should expand to a series of @code{asm} statements
5181 which will be compiled with GCC@. They go in a library function named
5182 @code{__transfer_from_trampoline}.
5184 If you need to avoid executing the ordinary prologue code of a compiled
5185 C function when you jump to the subroutine, you can do so by placing a
5186 special label of your own in the assembler code. Use one @code{asm}
5187 statement to generate an assembler label, and another to make the label
5188 global. Then trampolines can use that label to jump directly to your
5189 special assembler code.
5193 @section Implicit Calls to Library Routines
5194 @cindex library subroutine names
5195 @cindex @file{libgcc.a}
5197 @c prevent bad page break with this line
5198 Here is an explanation of implicit calls to library routines.
5200 @defmac DECLARE_LIBRARY_RENAMES
5201 This macro, if defined, should expand to a piece of C code that will get
5202 expanded when compiling functions for libgcc.a. It can be used to
5203 provide alternate names for GCC's internal library functions if there
5204 are ABI-mandated names that the compiler should provide.
5207 @findex set_optab_libfunc
5208 @findex init_one_libfunc
5209 @hook TARGET_INIT_LIBFUNCS
5210 This hook should declare additional library routines or rename
5211 existing ones, using the functions @code{set_optab_libfunc} and
5212 @code{init_one_libfunc} defined in @file{optabs.c}.
5213 @code{init_optabs} calls this macro after initializing all the normal
5216 The default is to do nothing. Most ports don't need to define this hook.
5219 @hook TARGET_LIBFUNC_GNU_PREFIX
5221 @defmac FLOAT_LIB_COMPARE_RETURNS_BOOL (@var{mode}, @var{comparison})
5222 This macro should return @code{true} if the library routine that
5223 implements the floating point comparison operator @var{comparison} in
5224 mode @var{mode} will return a boolean, and @var{false} if it will
5227 GCC's own floating point libraries return tristates from the
5228 comparison operators, so the default returns false always. Most ports
5229 don't need to define this macro.
5232 @defmac TARGET_LIB_INT_CMP_BIASED
5233 This macro should evaluate to @code{true} if the integer comparison
5234 functions (like @code{__cmpdi2}) return 0 to indicate that the first
5235 operand is smaller than the second, 1 to indicate that they are equal,
5236 and 2 to indicate that the first operand is greater than the second.
5237 If this macro evaluates to @code{false} the comparison functions return
5238 @minus{}1, 0, and 1 instead of 0, 1, and 2. If the target uses the routines
5239 in @file{libgcc.a}, you do not need to define this macro.
5242 @cindex @code{EDOM}, implicit usage
5245 The value of @code{EDOM} on the target machine, as a C integer constant
5246 expression. If you don't define this macro, GCC does not attempt to
5247 deposit the value of @code{EDOM} into @code{errno} directly. Look in
5248 @file{/usr/include/errno.h} to find the value of @code{EDOM} on your
5251 If you do not define @code{TARGET_EDOM}, then compiled code reports
5252 domain errors by calling the library function and letting it report the
5253 error. If mathematical functions on your system use @code{matherr} when
5254 there is an error, then you should leave @code{TARGET_EDOM} undefined so
5255 that @code{matherr} is used normally.
5258 @cindex @code{errno}, implicit usage
5259 @defmac GEN_ERRNO_RTX
5260 Define this macro as a C expression to create an rtl expression that
5261 refers to the global ``variable'' @code{errno}. (On certain systems,
5262 @code{errno} may not actually be a variable.) If you don't define this
5263 macro, a reasonable default is used.
5266 @cindex C99 math functions, implicit usage
5267 @defmac TARGET_C99_FUNCTIONS
5268 When this macro is nonzero, GCC will implicitly optimize @code{sin} calls into
5269 @code{sinf} and similarly for other functions defined by C99 standard. The
5270 default is zero because a number of existing systems lack support for these
5271 functions in their runtime so this macro needs to be redefined to one on
5272 systems that do support the C99 runtime.
5275 @cindex sincos math function, implicit usage
5276 @defmac TARGET_HAS_SINCOS
5277 When this macro is nonzero, GCC will implicitly optimize calls to @code{sin}
5278 and @code{cos} with the same argument to a call to @code{sincos}. The
5279 default is zero. The target has to provide the following functions:
5281 void sincos(double x, double *sin, double *cos);
5282 void sincosf(float x, float *sin, float *cos);
5283 void sincosl(long double x, long double *sin, long double *cos);
5287 @defmac NEXT_OBJC_RUNTIME
5288 Define this macro to generate code for Objective-C message sending using
5289 the calling convention of the NeXT system. This calling convention
5290 involves passing the object, the selector and the method arguments all
5291 at once to the method-lookup library function.
5293 The default calling convention passes just the object and the selector
5294 to the lookup function, which returns a pointer to the method.
5297 @node Addressing Modes
5298 @section Addressing Modes
5299 @cindex addressing modes
5301 @c prevent bad page break with this line
5302 This is about addressing modes.
5304 @defmac HAVE_PRE_INCREMENT
5305 @defmacx HAVE_PRE_DECREMENT
5306 @defmacx HAVE_POST_INCREMENT
5307 @defmacx HAVE_POST_DECREMENT
5308 A C expression that is nonzero if the machine supports pre-increment,
5309 pre-decrement, post-increment, or post-decrement addressing respectively.
5312 @defmac HAVE_PRE_MODIFY_DISP
5313 @defmacx HAVE_POST_MODIFY_DISP
5314 A C expression that is nonzero if the machine supports pre- or
5315 post-address side-effect generation involving constants other than
5316 the size of the memory operand.
5319 @defmac HAVE_PRE_MODIFY_REG
5320 @defmacx HAVE_POST_MODIFY_REG
5321 A C expression that is nonzero if the machine supports pre- or
5322 post-address side-effect generation involving a register displacement.
5325 @defmac CONSTANT_ADDRESS_P (@var{x})
5326 A C expression that is 1 if the RTX @var{x} is a constant which
5327 is a valid address. On most machines the default definition of
5328 @code{(CONSTANT_P (@var{x}) && GET_CODE (@var{x}) != CONST_DOUBLE)}
5329 is acceptable, but a few machines are more restrictive as to which
5330 constant addresses are supported.
5333 @defmac CONSTANT_P (@var{x})
5334 @code{CONSTANT_P}, which is defined by target-independent code,
5335 accepts integer-values expressions whose values are not explicitly
5336 known, such as @code{symbol_ref}, @code{label_ref}, and @code{high}
5337 expressions and @code{const} arithmetic expressions, in addition to
5338 @code{const_int} and @code{const_double} expressions.
5341 @defmac MAX_REGS_PER_ADDRESS
5342 A number, the maximum number of registers that can appear in a valid
5343 memory address. Note that it is up to you to specify a value equal to
5344 the maximum number that @code{TARGET_LEGITIMATE_ADDRESS_P} would ever
5348 @hook TARGET_LEGITIMATE_ADDRESS_P
5349 A function that returns whether @var{x} (an RTX) is a legitimate memory
5350 address on the target machine for a memory operand of mode @var{mode}.
5352 Legitimate addresses are defined in two variants: a strict variant and a
5353 non-strict one. The @var{strict} parameter chooses which variant is
5354 desired by the caller.
5356 The strict variant is used in the reload pass. It must be defined so
5357 that any pseudo-register that has not been allocated a hard register is
5358 considered a memory reference. This is because in contexts where some
5359 kind of register is required, a pseudo-register with no hard register
5360 must be rejected. For non-hard registers, the strict variant should look
5361 up the @code{reg_renumber} array; it should then proceed using the hard
5362 register number in the array, or treat the pseudo as a memory reference
5363 if the array holds @code{-1}.
5365 The non-strict variant is used in other passes. It must be defined to
5366 accept all pseudo-registers in every context where some kind of
5367 register is required.
5369 Normally, constant addresses which are the sum of a @code{symbol_ref}
5370 and an integer are stored inside a @code{const} RTX to mark them as
5371 constant. Therefore, there is no need to recognize such sums
5372 specifically as legitimate addresses. Normally you would simply
5373 recognize any @code{const} as legitimate.
5375 Usually @code{PRINT_OPERAND_ADDRESS} is not prepared to handle constant
5376 sums that are not marked with @code{const}. It assumes that a naked
5377 @code{plus} indicates indexing. If so, then you @emph{must} reject such
5378 naked constant sums as illegitimate addresses, so that none of them will
5379 be given to @code{PRINT_OPERAND_ADDRESS}.
5381 @cindex @code{TARGET_ENCODE_SECTION_INFO} and address validation
5382 On some machines, whether a symbolic address is legitimate depends on
5383 the section that the address refers to. On these machines, define the
5384 target hook @code{TARGET_ENCODE_SECTION_INFO} to store the information
5385 into the @code{symbol_ref}, and then check for it here. When you see a
5386 @code{const}, you will have to look inside it to find the
5387 @code{symbol_ref} in order to determine the section. @xref{Assembler
5390 @cindex @code{GO_IF_LEGITIMATE_ADDRESS}
5391 Some ports are still using a deprecated legacy substitute for
5392 this hook, the @code{GO_IF_LEGITIMATE_ADDRESS} macro. This macro
5396 #define GO_IF_LEGITIMATE_ADDRESS (@var{mode}, @var{x}, @var{label})
5400 and should @code{goto @var{label}} if the address @var{x} is a valid
5401 address on the target machine for a memory operand of mode @var{mode}.
5403 @findex REG_OK_STRICT
5404 Compiler source files that want to use the strict variant of this
5405 macro define the macro @code{REG_OK_STRICT}. You should use an
5406 @code{#ifdef REG_OK_STRICT} conditional to define the strict variant in
5407 that case and the non-strict variant otherwise.
5409 Using the hook is usually simpler because it limits the number of
5410 files that are recompiled when changes are made.
5413 @defmac TARGET_MEM_CONSTRAINT
5414 A single character to be used instead of the default @code{'m'}
5415 character for general memory addresses. This defines the constraint
5416 letter which matches the memory addresses accepted by
5417 @code{TARGET_LEGITIMATE_ADDRESS_P}. Define this macro if you want to
5418 support new address formats in your back end without changing the
5419 semantics of the @code{'m'} constraint. This is necessary in order to
5420 preserve functionality of inline assembly constructs using the
5421 @code{'m'} constraint.
5424 @defmac FIND_BASE_TERM (@var{x})
5425 A C expression to determine the base term of address @var{x},
5426 or to provide a simplified version of @var{x} from which @file{alias.c}
5427 can easily find the base term. This macro is used in only two places:
5428 @code{find_base_value} and @code{find_base_term} in @file{alias.c}.
5430 It is always safe for this macro to not be defined. It exists so
5431 that alias analysis can understand machine-dependent addresses.
5433 The typical use of this macro is to handle addresses containing
5434 a label_ref or symbol_ref within an UNSPEC@.
5437 @hook TARGET_LEGITIMIZE_ADDRESS
5438 This hook is given an invalid memory address @var{x} for an
5439 operand of mode @var{mode} and should try to return a valid memory
5442 @findex break_out_memory_refs
5443 @var{x} will always be the result of a call to @code{break_out_memory_refs},
5444 and @var{oldx} will be the operand that was given to that function to produce
5447 The code of the hook should not alter the substructure of
5448 @var{x}. If it transforms @var{x} into a more legitimate form, it
5449 should return the new @var{x}.
5451 It is not necessary for this hook to come up with a legitimate address.
5452 The compiler has standard ways of doing so in all cases. In fact, it
5453 is safe to omit this hook or make it return @var{x} if it cannot find
5454 a valid way to legitimize the address. But often a machine-dependent
5455 strategy can generate better code.
5458 @defmac LEGITIMIZE_RELOAD_ADDRESS (@var{x}, @var{mode}, @var{opnum}, @var{type}, @var{ind_levels}, @var{win})
5459 A C compound statement that attempts to replace @var{x}, which is an address
5460 that needs reloading, with a valid memory address for an operand of mode
5461 @var{mode}. @var{win} will be a C statement label elsewhere in the code.
5462 It is not necessary to define this macro, but it might be useful for
5463 performance reasons.
5465 For example, on the i386, it is sometimes possible to use a single
5466 reload register instead of two by reloading a sum of two pseudo
5467 registers into a register. On the other hand, for number of RISC
5468 processors offsets are limited so that often an intermediate address
5469 needs to be generated in order to address a stack slot. By defining
5470 @code{LEGITIMIZE_RELOAD_ADDRESS} appropriately, the intermediate addresses
5471 generated for adjacent some stack slots can be made identical, and thus
5474 @emph{Note}: This macro should be used with caution. It is necessary
5475 to know something of how reload works in order to effectively use this,
5476 and it is quite easy to produce macros that build in too much knowledge
5477 of reload internals.
5479 @emph{Note}: This macro must be able to reload an address created by a
5480 previous invocation of this macro. If it fails to handle such addresses
5481 then the compiler may generate incorrect code or abort.
5484 The macro definition should use @code{push_reload} to indicate parts that
5485 need reloading; @var{opnum}, @var{type} and @var{ind_levels} are usually
5486 suitable to be passed unaltered to @code{push_reload}.
5488 The code generated by this macro must not alter the substructure of
5489 @var{x}. If it transforms @var{x} into a more legitimate form, it
5490 should assign @var{x} (which will always be a C variable) a new value.
5491 This also applies to parts that you change indirectly by calling
5494 @findex strict_memory_address_p
5495 The macro definition may use @code{strict_memory_address_p} to test if
5496 the address has become legitimate.
5499 If you want to change only a part of @var{x}, one standard way of doing
5500 this is to use @code{copy_rtx}. Note, however, that it unshares only a
5501 single level of rtl. Thus, if the part to be changed is not at the
5502 top level, you'll need to replace first the top level.
5503 It is not necessary for this macro to come up with a legitimate
5504 address; but often a machine-dependent strategy can generate better code.
5507 @hook TARGET_MODE_DEPENDENT_ADDRESS_P
5508 This hook returns @code{true} if memory address @var{addr} can have
5509 different meanings depending on the machine mode of the memory
5510 reference it is used for or if the address is valid for some modes
5513 Autoincrement and autodecrement addresses typically have mode-dependent
5514 effects because the amount of the increment or decrement is the size
5515 of the operand being addressed. Some machines have other mode-dependent
5516 addresses. Many RISC machines have no mode-dependent addresses.
5518 You may assume that @var{addr} is a valid address for the machine.
5520 The default version of this hook returns @code{false}.
5523 @defmac GO_IF_MODE_DEPENDENT_ADDRESS (@var{addr}, @var{label})
5524 A C statement or compound statement with a conditional @code{goto
5525 @var{label};} executed if memory address @var{x} (an RTX) can have
5526 different meanings depending on the machine mode of the memory
5527 reference it is used for or if the address is valid for some modes
5530 Autoincrement and autodecrement addresses typically have mode-dependent
5531 effects because the amount of the increment or decrement is the size
5532 of the operand being addressed. Some machines have other mode-dependent
5533 addresses. Many RISC machines have no mode-dependent addresses.
5535 You may assume that @var{addr} is a valid address for the machine.
5537 These are obsolete macros, replaced by the
5538 @code{TARGET_MODE_DEPENDENT_ADDRESS_P} target hook.
5541 @hook TARGET_LEGITIMATE_CONSTANT_P
5542 This hook returns true if @var{x} is a legitimate constant for a
5543 @var{mode}-mode immediate operand on the target machine. You can assume that
5544 @var{x} satisfies @code{CONSTANT_P}, so you need not check this.
5546 The default definition returns true.
5549 @hook TARGET_DELEGITIMIZE_ADDRESS
5550 This hook is used to undo the possibly obfuscating effects of the
5551 @code{LEGITIMIZE_ADDRESS} and @code{LEGITIMIZE_RELOAD_ADDRESS} target
5552 macros. Some backend implementations of these macros wrap symbol
5553 references inside an @code{UNSPEC} rtx to represent PIC or similar
5554 addressing modes. This target hook allows GCC's optimizers to understand
5555 the semantics of these opaque @code{UNSPEC}s by converting them back
5556 into their original form.
5559 @hook TARGET_CANNOT_FORCE_CONST_MEM
5560 This hook should return true if @var{x} is of a form that cannot (or
5561 should not) be spilled to the constant pool. @var{mode} is the mode
5564 The default version of this hook returns false.
5566 The primary reason to define this hook is to prevent reload from
5567 deciding that a non-legitimate constant would be better reloaded
5568 from the constant pool instead of spilling and reloading a register
5569 holding the constant. This restriction is often true of addresses
5570 of TLS symbols for various targets.
5573 @hook TARGET_USE_BLOCKS_FOR_CONSTANT_P
5574 This hook should return true if pool entries for constant @var{x} can
5575 be placed in an @code{object_block} structure. @var{mode} is the mode
5578 The default version returns false for all constants.
5581 @hook TARGET_BUILTIN_RECIPROCAL
5582 This hook should return the DECL of a function that implements reciprocal of
5583 the builtin function with builtin function code @var{fn}, or
5584 @code{NULL_TREE} if such a function is not available. @var{md_fn} is true
5585 when @var{fn} is a code of a machine-dependent builtin function. When
5586 @var{sqrt} is true, additional optimizations that apply only to the reciprocal
5587 of a square root function are performed, and only reciprocals of @code{sqrt}
5591 @hook TARGET_VECTORIZE_BUILTIN_MASK_FOR_LOAD
5592 This hook should return the DECL of a function @var{f} that given an
5593 address @var{addr} as an argument returns a mask @var{m} that can be
5594 used to extract from two vectors the relevant data that resides in
5595 @var{addr} in case @var{addr} is not properly aligned.
5597 The autovectorizer, when vectorizing a load operation from an address
5598 @var{addr} that may be unaligned, will generate two vector loads from
5599 the two aligned addresses around @var{addr}. It then generates a
5600 @code{REALIGN_LOAD} operation to extract the relevant data from the
5601 two loaded vectors. The first two arguments to @code{REALIGN_LOAD},
5602 @var{v1} and @var{v2}, are the two vectors, each of size @var{VS}, and
5603 the third argument, @var{OFF}, defines how the data will be extracted
5604 from these two vectors: if @var{OFF} is 0, then the returned vector is
5605 @var{v2}; otherwise, the returned vector is composed from the last
5606 @var{VS}-@var{OFF} elements of @var{v1} concatenated to the first
5607 @var{OFF} elements of @var{v2}.
5609 If this hook is defined, the autovectorizer will generate a call
5610 to @var{f} (using the DECL tree that this hook returns) and will
5611 use the return value of @var{f} as the argument @var{OFF} to
5612 @code{REALIGN_LOAD}. Therefore, the mask @var{m} returned by @var{f}
5613 should comply with the semantics expected by @code{REALIGN_LOAD}
5615 If this hook is not defined, then @var{addr} will be used as
5616 the argument @var{OFF} to @code{REALIGN_LOAD}, in which case the low
5617 log2(@var{VS}) @minus{} 1 bits of @var{addr} will be considered.
5620 @hook TARGET_VECTORIZE_BUILTIN_MUL_WIDEN_EVEN
5621 This hook should return the DECL of a function @var{f} that implements
5622 widening multiplication of the even elements of two input vectors of type @var{x}.
5624 If this hook is defined, the autovectorizer will use it along with the
5625 @code{TARGET_VECTORIZE_BUILTIN_MUL_WIDEN_ODD} target hook when vectorizing
5626 widening multiplication in cases that the order of the results does not have to be
5627 preserved (e.g.@: used only by a reduction computation). Otherwise, the
5628 @code{widen_mult_hi/lo} idioms will be used.
5631 @hook TARGET_VECTORIZE_BUILTIN_MUL_WIDEN_ODD
5632 This hook should return the DECL of a function @var{f} that implements
5633 widening multiplication of the odd elements of two input vectors of type @var{x}.
5635 If this hook is defined, the autovectorizer will use it along with the
5636 @code{TARGET_VECTORIZE_BUILTIN_MUL_WIDEN_EVEN} target hook when vectorizing
5637 widening multiplication in cases that the order of the results does not have to be
5638 preserved (e.g.@: used only by a reduction computation). Otherwise, the
5639 @code{widen_mult_hi/lo} idioms will be used.
5642 @hook TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST
5643 Returns cost of different scalar or vector statements for vectorization cost model.
5644 For vector memory operations the cost may depend on type (@var{vectype}) and
5645 misalignment value (@var{misalign}).
5648 @hook TARGET_VECTORIZE_VECTOR_ALIGNMENT_REACHABLE
5649 Return true if vector alignment is reachable (by peeling N iterations) for the given type.
5652 @hook TARGET_VECTORIZE_VEC_PERM_CONST_OK
5653 Return true if a vector created for @code{vec_perm_const} is valid.
5656 @hook TARGET_VECTORIZE_BUILTIN_CONVERSION
5657 This hook should return the DECL of a function that implements conversion of the
5658 input vector of type @var{src_type} to type @var{dest_type}.
5659 The value of @var{code} is one of the enumerators in @code{enum tree_code} and
5660 specifies how the conversion is to be applied
5661 (truncation, rounding, etc.).
5663 If this hook is defined, the autovectorizer will use the
5664 @code{TARGET_VECTORIZE_BUILTIN_CONVERSION} target hook when vectorizing
5665 conversion. Otherwise, it will return @code{NULL_TREE}.
5668 @hook TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION
5669 This hook should return the decl of a function that implements the
5670 vectorized variant of the builtin function with builtin function code
5671 @var{code} or @code{NULL_TREE} if such a function is not available.
5672 The value of @var{fndecl} is the builtin function declaration. The
5673 return type of the vectorized function shall be of vector type
5674 @var{vec_type_out} and the argument types should be @var{vec_type_in}.
5677 @hook TARGET_VECTORIZE_SUPPORT_VECTOR_MISALIGNMENT
5678 This hook should return true if the target supports misaligned vector
5679 store/load of a specific factor denoted in the @var{misalignment}
5680 parameter. The vector store/load should be of machine mode @var{mode} and
5681 the elements in the vectors should be of type @var{type}. @var{is_packed}
5682 parameter is true if the memory access is defined in a packed struct.
5685 @hook TARGET_VECTORIZE_PREFERRED_SIMD_MODE
5686 This hook should return the preferred mode for vectorizing scalar
5687 mode @var{mode}. The default is
5688 equal to @code{word_mode}, because the vectorizer can do some
5689 transformations even in absence of specialized @acronym{SIMD} hardware.
5692 @hook TARGET_VECTORIZE_AUTOVECTORIZE_VECTOR_SIZES
5693 This hook should return a mask of sizes that should be iterated over
5694 after trying to autovectorize using the vector size derived from the
5695 mode returned by @code{TARGET_VECTORIZE_PREFERRED_SIMD_MODE}.
5696 The default is zero which means to not iterate over other vector sizes.
5699 @hook TARGET_VECTORIZE_BUILTIN_TM_LOAD
5701 @hook TARGET_VECTORIZE_BUILTIN_TM_STORE
5703 @hook TARGET_VECTORIZE_BUILTIN_GATHER
5704 Target builtin that implements vector gather operation. @var{mem_vectype}
5705 is the vector type of the load and @var{index_type} is scalar type of
5706 the index, scaled by @var{scale}.
5707 The default is @code{NULL_TREE} which means to not vectorize gather
5711 @node Anchored Addresses
5712 @section Anchored Addresses
5713 @cindex anchored addresses
5714 @cindex @option{-fsection-anchors}
5716 GCC usually addresses every static object as a separate entity.
5717 For example, if we have:
5721 int foo (void) @{ return a + b + c; @}
5724 the code for @code{foo} will usually calculate three separate symbolic
5725 addresses: those of @code{a}, @code{b} and @code{c}. On some targets,
5726 it would be better to calculate just one symbolic address and access
5727 the three variables relative to it. The equivalent pseudocode would
5733 register int *xr = &x;
5734 return xr[&a - &x] + xr[&b - &x] + xr[&c - &x];
5738 (which isn't valid C). We refer to shared addresses like @code{x} as
5739 ``section anchors''. Their use is controlled by @option{-fsection-anchors}.
5741 The hooks below describe the target properties that GCC needs to know
5742 in order to make effective use of section anchors. It won't use
5743 section anchors at all unless either @code{TARGET_MIN_ANCHOR_OFFSET}
5744 or @code{TARGET_MAX_ANCHOR_OFFSET} is set to a nonzero value.
5746 @hook TARGET_MIN_ANCHOR_OFFSET
5747 The minimum offset that should be applied to a section anchor.
5748 On most targets, it should be the smallest offset that can be
5749 applied to a base register while still giving a legitimate address
5750 for every mode. The default value is 0.
5753 @hook TARGET_MAX_ANCHOR_OFFSET
5754 Like @code{TARGET_MIN_ANCHOR_OFFSET}, but the maximum (inclusive)
5755 offset that should be applied to section anchors. The default
5759 @hook TARGET_ASM_OUTPUT_ANCHOR
5760 Write the assembly code to define section anchor @var{x}, which is a
5761 @code{SYMBOL_REF} for which @samp{SYMBOL_REF_ANCHOR_P (@var{x})} is true.
5762 The hook is called with the assembly output position set to the beginning
5763 of @code{SYMBOL_REF_BLOCK (@var{x})}.
5765 If @code{ASM_OUTPUT_DEF} is available, the hook's default definition uses
5766 it to define the symbol as @samp{. + SYMBOL_REF_BLOCK_OFFSET (@var{x})}.
5767 If @code{ASM_OUTPUT_DEF} is not available, the hook's default definition
5768 is @code{NULL}, which disables the use of section anchors altogether.
5771 @hook TARGET_USE_ANCHORS_FOR_SYMBOL_P
5772 Return true if GCC should attempt to use anchors to access @code{SYMBOL_REF}
5773 @var{x}. You can assume @samp{SYMBOL_REF_HAS_BLOCK_INFO_P (@var{x})} and
5774 @samp{!SYMBOL_REF_ANCHOR_P (@var{x})}.
5776 The default version is correct for most targets, but you might need to
5777 intercept this hook to handle things like target-specific attributes
5778 or target-specific sections.
5781 @node Condition Code
5782 @section Condition Code Status
5783 @cindex condition code status
5785 The macros in this section can be split in two families, according to the
5786 two ways of representing condition codes in GCC.
5788 The first representation is the so called @code{(cc0)} representation
5789 (@pxref{Jump Patterns}), where all instructions can have an implicit
5790 clobber of the condition codes. The second is the condition code
5791 register representation, which provides better schedulability for
5792 architectures that do have a condition code register, but on which
5793 most instructions do not affect it. The latter category includes
5796 The implicit clobbering poses a strong restriction on the placement of
5797 the definition and use of the condition code, which need to be in adjacent
5798 insns for machines using @code{(cc0)}. This can prevent important
5799 optimizations on some machines. For example, on the IBM RS/6000, there
5800 is a delay for taken branches unless the condition code register is set
5801 three instructions earlier than the conditional branch. The instruction
5802 scheduler cannot perform this optimization if it is not permitted to
5803 separate the definition and use of the condition code register.
5805 For this reason, it is possible and suggested to use a register to
5806 represent the condition code for new ports. If there is a specific
5807 condition code register in the machine, use a hard register. If the
5808 condition code or comparison result can be placed in any general register,
5809 or if there are multiple condition registers, use a pseudo register.
5810 Registers used to store the condition code value will usually have a mode
5811 that is in class @code{MODE_CC}.
5813 Alternatively, you can use @code{BImode} if the comparison operator is
5814 specified already in the compare instruction. In this case, you are not
5815 interested in most macros in this section.
5818 * CC0 Condition Codes:: Old style representation of condition codes.
5819 * MODE_CC Condition Codes:: Modern representation of condition codes.
5820 * Cond Exec Macros:: Macros to control conditional execution.
5823 @node CC0 Condition Codes
5824 @subsection Representation of condition codes using @code{(cc0)}
5828 The file @file{conditions.h} defines a variable @code{cc_status} to
5829 describe how the condition code was computed (in case the interpretation of
5830 the condition code depends on the instruction that it was set by). This
5831 variable contains the RTL expressions on which the condition code is
5832 currently based, and several standard flags.
5834 Sometimes additional machine-specific flags must be defined in the machine
5835 description header file. It can also add additional machine-specific
5836 information by defining @code{CC_STATUS_MDEP}.
5838 @defmac CC_STATUS_MDEP
5839 C code for a data type which is used for declaring the @code{mdep}
5840 component of @code{cc_status}. It defaults to @code{int}.
5842 This macro is not used on machines that do not use @code{cc0}.
5845 @defmac CC_STATUS_MDEP_INIT
5846 A C expression to initialize the @code{mdep} field to ``empty''.
5847 The default definition does nothing, since most machines don't use
5848 the field anyway. If you want to use the field, you should probably
5849 define this macro to initialize it.
5851 This macro is not used on machines that do not use @code{cc0}.
5854 @defmac NOTICE_UPDATE_CC (@var{exp}, @var{insn})
5855 A C compound statement to set the components of @code{cc_status}
5856 appropriately for an insn @var{insn} whose body is @var{exp}. It is
5857 this macro's responsibility to recognize insns that set the condition
5858 code as a byproduct of other activity as well as those that explicitly
5861 This macro is not used on machines that do not use @code{cc0}.
5863 If there are insns that do not set the condition code but do alter
5864 other machine registers, this macro must check to see whether they
5865 invalidate the expressions that the condition code is recorded as
5866 reflecting. For example, on the 68000, insns that store in address
5867 registers do not set the condition code, which means that usually
5868 @code{NOTICE_UPDATE_CC} can leave @code{cc_status} unaltered for such
5869 insns. But suppose that the previous insn set the condition code
5870 based on location @samp{a4@@(102)} and the current insn stores a new
5871 value in @samp{a4}. Although the condition code is not changed by
5872 this, it will no longer be true that it reflects the contents of
5873 @samp{a4@@(102)}. Therefore, @code{NOTICE_UPDATE_CC} must alter
5874 @code{cc_status} in this case to say that nothing is known about the
5875 condition code value.
5877 The definition of @code{NOTICE_UPDATE_CC} must be prepared to deal
5878 with the results of peephole optimization: insns whose patterns are
5879 @code{parallel} RTXs containing various @code{reg}, @code{mem} or
5880 constants which are just the operands. The RTL structure of these
5881 insns is not sufficient to indicate what the insns actually do. What
5882 @code{NOTICE_UPDATE_CC} should do when it sees one is just to run
5883 @code{CC_STATUS_INIT}.
5885 A possible definition of @code{NOTICE_UPDATE_CC} is to call a function
5886 that looks at an attribute (@pxref{Insn Attributes}) named, for example,
5887 @samp{cc}. This avoids having detailed information about patterns in
5888 two places, the @file{md} file and in @code{NOTICE_UPDATE_CC}.
5891 @node MODE_CC Condition Codes
5892 @subsection Representation of condition codes using registers
5896 @defmac SELECT_CC_MODE (@var{op}, @var{x}, @var{y})
5897 On many machines, the condition code may be produced by other instructions
5898 than compares, for example the branch can use directly the condition
5899 code set by a subtract instruction. However, on some machines
5900 when the condition code is set this way some bits (such as the overflow
5901 bit) are not set in the same way as a test instruction, so that a different
5902 branch instruction must be used for some conditional branches. When
5903 this happens, use the machine mode of the condition code register to
5904 record different formats of the condition code register. Modes can
5905 also be used to record which compare instruction (e.g. a signed or an
5906 unsigned comparison) produced the condition codes.
5908 If other modes than @code{CCmode} are required, add them to
5909 @file{@var{machine}-modes.def} and define @code{SELECT_CC_MODE} to choose
5910 a mode given an operand of a compare. This is needed because the modes
5911 have to be chosen not only during RTL generation but also, for example,
5912 by instruction combination. The result of @code{SELECT_CC_MODE} should
5913 be consistent with the mode used in the patterns; for example to support
5914 the case of the add on the SPARC discussed above, we have the pattern
5918 [(set (reg:CC_NOOV 0)
5920 (plus:SI (match_operand:SI 0 "register_operand" "%r")
5921 (match_operand:SI 1 "arith_operand" "rI"))
5928 together with a @code{SELECT_CC_MODE} that returns @code{CC_NOOVmode}
5929 for comparisons whose argument is a @code{plus}:
5932 #define SELECT_CC_MODE(OP,X,Y) \
5933 (GET_MODE_CLASS (GET_MODE (X)) == MODE_FLOAT \
5934 ? ((OP == EQ || OP == NE) ? CCFPmode : CCFPEmode) \
5935 : ((GET_CODE (X) == PLUS || GET_CODE (X) == MINUS \
5936 || GET_CODE (X) == NEG) \
5937 ? CC_NOOVmode : CCmode))
5940 Another reason to use modes is to retain information on which operands
5941 were used by the comparison; see @code{REVERSIBLE_CC_MODE} later in
5944 You should define this macro if and only if you define extra CC modes
5945 in @file{@var{machine}-modes.def}.
5948 @defmac CANONICALIZE_COMPARISON (@var{code}, @var{op0}, @var{op1})
5949 On some machines not all possible comparisons are defined, but you can
5950 convert an invalid comparison into a valid one. For example, the Alpha
5951 does not have a @code{GT} comparison, but you can use an @code{LT}
5952 comparison instead and swap the order of the operands.
5954 On such machines, define this macro to be a C statement to do any
5955 required conversions. @var{code} is the initial comparison code
5956 and @var{op0} and @var{op1} are the left and right operands of the
5957 comparison, respectively. You should modify @var{code}, @var{op0}, and
5958 @var{op1} as required.
5960 GCC will not assume that the comparison resulting from this macro is
5961 valid but will see if the resulting insn matches a pattern in the
5964 You need not define this macro if it would never change the comparison
5968 @defmac REVERSIBLE_CC_MODE (@var{mode})
5969 A C expression whose value is one if it is always safe to reverse a
5970 comparison whose mode is @var{mode}. If @code{SELECT_CC_MODE}
5971 can ever return @var{mode} for a floating-point inequality comparison,
5972 then @code{REVERSIBLE_CC_MODE (@var{mode})} must be zero.
5974 You need not define this macro if it would always returns zero or if the
5975 floating-point format is anything other than @code{IEEE_FLOAT_FORMAT}.
5976 For example, here is the definition used on the SPARC, where floating-point
5977 inequality comparisons are always given @code{CCFPEmode}:
5980 #define REVERSIBLE_CC_MODE(MODE) ((MODE) != CCFPEmode)
5984 @defmac REVERSE_CONDITION (@var{code}, @var{mode})
5985 A C expression whose value is reversed condition code of the @var{code} for
5986 comparison done in CC_MODE @var{mode}. The macro is used only in case
5987 @code{REVERSIBLE_CC_MODE (@var{mode})} is nonzero. Define this macro in case
5988 machine has some non-standard way how to reverse certain conditionals. For
5989 instance in case all floating point conditions are non-trapping, compiler may
5990 freely convert unordered compares to ordered one. Then definition may look
5994 #define REVERSE_CONDITION(CODE, MODE) \
5995 ((MODE) != CCFPmode ? reverse_condition (CODE) \
5996 : reverse_condition_maybe_unordered (CODE))
6000 @hook TARGET_FIXED_CONDITION_CODE_REGS
6001 On targets which do not use @code{(cc0)}, and which use a hard
6002 register rather than a pseudo-register to hold condition codes, the
6003 regular CSE passes are often not able to identify cases in which the
6004 hard register is set to a common value. Use this hook to enable a
6005 small pass which optimizes such cases. This hook should return true
6006 to enable this pass, and it should set the integers to which its
6007 arguments point to the hard register numbers used for condition codes.
6008 When there is only one such register, as is true on most systems, the
6009 integer pointed to by @var{p2} should be set to
6010 @code{INVALID_REGNUM}.
6012 The default version of this hook returns false.
6015 @hook TARGET_CC_MODES_COMPATIBLE
6016 On targets which use multiple condition code modes in class
6017 @code{MODE_CC}, it is sometimes the case that a comparison can be
6018 validly done in more than one mode. On such a system, define this
6019 target hook to take two mode arguments and to return a mode in which
6020 both comparisons may be validly done. If there is no such mode,
6021 return @code{VOIDmode}.
6023 The default version of this hook checks whether the modes are the
6024 same. If they are, it returns that mode. If they are different, it
6025 returns @code{VOIDmode}.
6028 @node Cond Exec Macros
6029 @subsection Macros to control conditional execution
6030 @findex conditional execution
6033 There is one macro that may need to be defined for targets
6034 supporting conditional execution, independent of how they
6035 represent conditional branches.
6037 @defmac REVERSE_CONDEXEC_PREDICATES_P (@var{op1}, @var{op2})
6038 A C expression that returns true if the conditional execution predicate
6039 @var{op1}, a comparison operation, is the inverse of @var{op2} and vice
6040 versa. Define this to return 0 if the target has conditional execution
6041 predicates that cannot be reversed safely. There is no need to validate
6042 that the arguments of op1 and op2 are the same, this is done separately.
6043 If no expansion is specified, this macro is defined as follows:
6046 #define REVERSE_CONDEXEC_PREDICATES_P (x, y) \
6047 (GET_CODE ((x)) == reversed_comparison_code ((y), NULL))
6052 @section Describing Relative Costs of Operations
6053 @cindex costs of instructions
6054 @cindex relative costs
6055 @cindex speed of instructions
6057 These macros let you describe the relative speed of various operations
6058 on the target machine.
6060 @defmac REGISTER_MOVE_COST (@var{mode}, @var{from}, @var{to})
6061 A C expression for the cost of moving data of mode @var{mode} from a
6062 register in class @var{from} to one in class @var{to}. The classes are
6063 expressed using the enumeration values such as @code{GENERAL_REGS}. A
6064 value of 2 is the default; other values are interpreted relative to
6067 It is not required that the cost always equal 2 when @var{from} is the
6068 same as @var{to}; on some machines it is expensive to move between
6069 registers if they are not general registers.
6071 If reload sees an insn consisting of a single @code{set} between two
6072 hard registers, and if @code{REGISTER_MOVE_COST} applied to their
6073 classes returns a value of 2, reload does not check to ensure that the
6074 constraints of the insn are met. Setting a cost of other than 2 will
6075 allow reload to verify that the constraints are met. You should do this
6076 if the @samp{mov@var{m}} pattern's constraints do not allow such copying.
6078 These macros are obsolete, new ports should use the target hook
6079 @code{TARGET_REGISTER_MOVE_COST} instead.
6082 @hook TARGET_REGISTER_MOVE_COST
6083 This target hook should return the cost of moving data of mode @var{mode}
6084 from a register in class @var{from} to one in class @var{to}. The classes
6085 are expressed using the enumeration values such as @code{GENERAL_REGS}.
6086 A value of 2 is the default; other values are interpreted relative to
6089 It is not required that the cost always equal 2 when @var{from} is the
6090 same as @var{to}; on some machines it is expensive to move between
6091 registers if they are not general registers.
6093 If reload sees an insn consisting of a single @code{set} between two
6094 hard registers, and if @code{TARGET_REGISTER_MOVE_COST} applied to their
6095 classes returns a value of 2, reload does not check to ensure that the
6096 constraints of the insn are met. Setting a cost of other than 2 will
6097 allow reload to verify that the constraints are met. You should do this
6098 if the @samp{mov@var{m}} pattern's constraints do not allow such copying.
6100 The default version of this function returns 2.
6103 @defmac MEMORY_MOVE_COST (@var{mode}, @var{class}, @var{in})
6104 A C expression for the cost of moving data of mode @var{mode} between a
6105 register of class @var{class} and memory; @var{in} is zero if the value
6106 is to be written to memory, nonzero if it is to be read in. This cost
6107 is relative to those in @code{REGISTER_MOVE_COST}. If moving between
6108 registers and memory is more expensive than between two registers, you
6109 should define this macro to express the relative cost.
6111 If you do not define this macro, GCC uses a default cost of 4 plus
6112 the cost of copying via a secondary reload register, if one is
6113 needed. If your machine requires a secondary reload register to copy
6114 between memory and a register of @var{class} but the reload mechanism is
6115 more complex than copying via an intermediate, define this macro to
6116 reflect the actual cost of the move.
6118 GCC defines the function @code{memory_move_secondary_cost} if
6119 secondary reloads are needed. It computes the costs due to copying via
6120 a secondary register. If your machine copies from memory using a
6121 secondary register in the conventional way but the default base value of
6122 4 is not correct for your machine, define this macro to add some other
6123 value to the result of that function. The arguments to that function
6124 are the same as to this macro.
6126 These macros are obsolete, new ports should use the target hook
6127 @code{TARGET_MEMORY_MOVE_COST} instead.
6130 @hook TARGET_MEMORY_MOVE_COST
6131 This target hook should return the cost of moving data of mode @var{mode}
6132 between a register of class @var{rclass} and memory; @var{in} is @code{false}
6133 if the value is to be written to memory, @code{true} if it is to be read in.
6134 This cost is relative to those in @code{TARGET_REGISTER_MOVE_COST}.
6135 If moving between registers and memory is more expensive than between two
6136 registers, you should add this target hook to express the relative cost.
6138 If you do not add this target hook, GCC uses a default cost of 4 plus
6139 the cost of copying via a secondary reload register, if one is
6140 needed. If your machine requires a secondary reload register to copy
6141 between memory and a register of @var{rclass} but the reload mechanism is
6142 more complex than copying via an intermediate, use this target hook to
6143 reflect the actual cost of the move.
6145 GCC defines the function @code{memory_move_secondary_cost} if
6146 secondary reloads are needed. It computes the costs due to copying via
6147 a secondary register. If your machine copies from memory using a
6148 secondary register in the conventional way but the default base value of
6149 4 is not correct for your machine, use this target hook to add some other
6150 value to the result of that function. The arguments to that function
6151 are the same as to this target hook.
6154 @defmac BRANCH_COST (@var{speed_p}, @var{predictable_p})
6155 A C expression for the cost of a branch instruction. A value of 1 is
6156 the default; other values are interpreted relative to that. Parameter
6157 @var{speed_p} is true when the branch in question should be optimized
6158 for speed. When it is false, @code{BRANCH_COST} should return a value
6159 optimal for code size rather than performance. @var{predictable_p} is
6160 true for well-predicted branches. On many architectures the
6161 @code{BRANCH_COST} can be reduced then.
6164 Here are additional macros which do not specify precise relative costs,
6165 but only that certain actions are more expensive than GCC would
6168 @defmac SLOW_BYTE_ACCESS
6169 Define this macro as a C expression which is nonzero if accessing less
6170 than a word of memory (i.e.@: a @code{char} or a @code{short}) is no
6171 faster than accessing a word of memory, i.e., if such access
6172 require more than one instruction or if there is no difference in cost
6173 between byte and (aligned) word loads.
6175 When this macro is not defined, the compiler will access a field by
6176 finding the smallest containing object; when it is defined, a fullword
6177 load will be used if alignment permits. Unless bytes accesses are
6178 faster than word accesses, using word accesses is preferable since it
6179 may eliminate subsequent memory access if subsequent accesses occur to
6180 other fields in the same word of the structure, but to different bytes.
6183 @defmac SLOW_UNALIGNED_ACCESS (@var{mode}, @var{alignment})
6184 Define this macro to be the value 1 if memory accesses described by the
6185 @var{mode} and @var{alignment} parameters have a cost many times greater
6186 than aligned accesses, for example if they are emulated in a trap
6189 When this macro is nonzero, the compiler will act as if
6190 @code{STRICT_ALIGNMENT} were nonzero when generating code for block
6191 moves. This can cause significantly more instructions to be produced.
6192 Therefore, do not set this macro nonzero if unaligned accesses only add a
6193 cycle or two to the time for a memory access.
6195 If the value of this macro is always zero, it need not be defined. If
6196 this macro is defined, it should produce a nonzero value when
6197 @code{STRICT_ALIGNMENT} is nonzero.
6200 @defmac MOVE_RATIO (@var{speed})
6201 The threshold of number of scalar memory-to-memory move insns, @emph{below}
6202 which a sequence of insns should be generated instead of a
6203 string move insn or a library call. Increasing the value will always
6204 make code faster, but eventually incurs high cost in increased code size.
6206 Note that on machines where the corresponding move insn is a
6207 @code{define_expand} that emits a sequence of insns, this macro counts
6208 the number of such sequences.
6210 The parameter @var{speed} is true if the code is currently being
6211 optimized for speed rather than size.
6213 If you don't define this, a reasonable default is used.
6216 @defmac MOVE_BY_PIECES_P (@var{size}, @var{alignment})
6217 A C expression used to determine whether @code{move_by_pieces} will be used to
6218 copy a chunk of memory, or whether some other block move mechanism
6219 will be used. Defaults to 1 if @code{move_by_pieces_ninsns} returns less
6220 than @code{MOVE_RATIO}.
6223 @defmac MOVE_MAX_PIECES
6224 A C expression used by @code{move_by_pieces} to determine the largest unit
6225 a load or store used to copy memory is. Defaults to @code{MOVE_MAX}.
6228 @defmac CLEAR_RATIO (@var{speed})
6229 The threshold of number of scalar move insns, @emph{below} which a sequence
6230 of insns should be generated to clear memory instead of a string clear insn
6231 or a library call. Increasing the value will always make code faster, but
6232 eventually incurs high cost in increased code size.
6234 The parameter @var{speed} is true if the code is currently being
6235 optimized for speed rather than size.
6237 If you don't define this, a reasonable default is used.
6240 @defmac CLEAR_BY_PIECES_P (@var{size}, @var{alignment})
6241 A C expression used to determine whether @code{clear_by_pieces} will be used
6242 to clear a chunk of memory, or whether some other block clear mechanism
6243 will be used. Defaults to 1 if @code{move_by_pieces_ninsns} returns less
6244 than @code{CLEAR_RATIO}.
6247 @defmac SET_RATIO (@var{speed})
6248 The threshold of number of scalar move insns, @emph{below} which a sequence
6249 of insns should be generated to set memory to a constant value, instead of
6250 a block set insn or a library call.
6251 Increasing the value will always make code faster, but
6252 eventually incurs high cost in increased code size.
6254 The parameter @var{speed} is true if the code is currently being
6255 optimized for speed rather than size.
6257 If you don't define this, it defaults to the value of @code{MOVE_RATIO}.
6260 @defmac SET_BY_PIECES_P (@var{size}, @var{alignment})
6261 A C expression used to determine whether @code{store_by_pieces} will be
6262 used to set a chunk of memory to a constant value, or whether some
6263 other mechanism will be used. Used by @code{__builtin_memset} when
6264 storing values other than constant zero.
6265 Defaults to 1 if @code{move_by_pieces_ninsns} returns less
6266 than @code{SET_RATIO}.
6269 @defmac STORE_BY_PIECES_P (@var{size}, @var{alignment})
6270 A C expression used to determine whether @code{store_by_pieces} will be
6271 used to set a chunk of memory to a constant string value, or whether some
6272 other mechanism will be used. Used by @code{__builtin_strcpy} when
6273 called with a constant source string.
6274 Defaults to 1 if @code{move_by_pieces_ninsns} returns less
6275 than @code{MOVE_RATIO}.
6278 @defmac USE_LOAD_POST_INCREMENT (@var{mode})
6279 A C expression used to determine whether a load postincrement is a good
6280 thing to use for a given mode. Defaults to the value of
6281 @code{HAVE_POST_INCREMENT}.
6284 @defmac USE_LOAD_POST_DECREMENT (@var{mode})
6285 A C expression used to determine whether a load postdecrement is a good
6286 thing to use for a given mode. Defaults to the value of
6287 @code{HAVE_POST_DECREMENT}.
6290 @defmac USE_LOAD_PRE_INCREMENT (@var{mode})
6291 A C expression used to determine whether a load preincrement is a good
6292 thing to use for a given mode. Defaults to the value of
6293 @code{HAVE_PRE_INCREMENT}.
6296 @defmac USE_LOAD_PRE_DECREMENT (@var{mode})
6297 A C expression used to determine whether a load predecrement is a good
6298 thing to use for a given mode. Defaults to the value of
6299 @code{HAVE_PRE_DECREMENT}.
6302 @defmac USE_STORE_POST_INCREMENT (@var{mode})
6303 A C expression used to determine whether a store postincrement is a good
6304 thing to use for a given mode. Defaults to the value of
6305 @code{HAVE_POST_INCREMENT}.
6308 @defmac USE_STORE_POST_DECREMENT (@var{mode})
6309 A C expression used to determine whether a store postdecrement is a good
6310 thing to use for a given mode. Defaults to the value of
6311 @code{HAVE_POST_DECREMENT}.
6314 @defmac USE_STORE_PRE_INCREMENT (@var{mode})
6315 This macro is used to determine whether a store preincrement is a good
6316 thing to use for a given mode. Defaults to the value of
6317 @code{HAVE_PRE_INCREMENT}.
6320 @defmac USE_STORE_PRE_DECREMENT (@var{mode})
6321 This macro is used to determine whether a store predecrement is a good
6322 thing to use for a given mode. Defaults to the value of
6323 @code{HAVE_PRE_DECREMENT}.
6326 @defmac NO_FUNCTION_CSE
6327 Define this macro if it is as good or better to call a constant
6328 function address than to call an address kept in a register.
6331 @defmac RANGE_TEST_NON_SHORT_CIRCUIT
6332 Define this macro if a non-short-circuit operation produced by
6333 @samp{fold_range_test ()} is optimal. This macro defaults to true if
6334 @code{BRANCH_COST} is greater than or equal to the value 2.
6337 @hook TARGET_RTX_COSTS
6338 This target hook describes the relative costs of RTL expressions.
6340 The cost may depend on the precise form of the expression, which is
6341 available for examination in @var{x}, and the fact that @var{x} appears
6342 as operand @var{opno} of an expression with rtx code @var{outer_code}.
6343 That is, the hook can assume that there is some rtx @var{y} such
6344 that @samp{GET_CODE (@var{y}) == @var{outer_code}} and such that
6345 either (a) @samp{XEXP (@var{y}, @var{opno}) == @var{x}} or
6346 (b) @samp{XVEC (@var{y}, @var{opno})} contains @var{x}.
6348 @var{code} is @var{x}'s expression code---redundant, since it can be
6349 obtained with @code{GET_CODE (@var{x})}.
6351 In implementing this hook, you can use the construct
6352 @code{COSTS_N_INSNS (@var{n})} to specify a cost equal to @var{n} fast
6355 On entry to the hook, @code{*@var{total}} contains a default estimate
6356 for the cost of the expression. The hook should modify this value as
6357 necessary. Traditionally, the default costs are @code{COSTS_N_INSNS (5)}
6358 for multiplications, @code{COSTS_N_INSNS (7)} for division and modulus
6359 operations, and @code{COSTS_N_INSNS (1)} for all other operations.
6361 When optimizing for code size, i.e.@: when @code{speed} is
6362 false, this target hook should be used to estimate the relative
6363 size cost of an expression, again relative to @code{COSTS_N_INSNS}.
6365 The hook returns true when all subexpressions of @var{x} have been
6366 processed, and false when @code{rtx_cost} should recurse.
6369 @hook TARGET_ADDRESS_COST
6370 This hook computes the cost of an addressing mode that contains
6371 @var{address}. If not defined, the cost is computed from
6372 the @var{address} expression and the @code{TARGET_RTX_COST} hook.
6374 For most CISC machines, the default cost is a good approximation of the
6375 true cost of the addressing mode. However, on RISC machines, all
6376 instructions normally have the same length and execution time. Hence
6377 all addresses will have equal costs.
6379 In cases where more than one form of an address is known, the form with
6380 the lowest cost will be used. If multiple forms have the same, lowest,
6381 cost, the one that is the most complex will be used.
6383 For example, suppose an address that is equal to the sum of a register
6384 and a constant is used twice in the same basic block. When this macro
6385 is not defined, the address will be computed in a register and memory
6386 references will be indirect through that register. On machines where
6387 the cost of the addressing mode containing the sum is no higher than
6388 that of a simple indirect reference, this will produce an additional
6389 instruction and possibly require an additional register. Proper
6390 specification of this macro eliminates this overhead for such machines.
6392 This hook is never called with an invalid address.
6394 On machines where an address involving more than one register is as
6395 cheap as an address computation involving only one register, defining
6396 @code{TARGET_ADDRESS_COST} to reflect this can cause two registers to
6397 be live over a region of code where only one would have been if
6398 @code{TARGET_ADDRESS_COST} were not defined in that manner. This effect
6399 should be considered in the definition of this macro. Equivalent costs
6400 should probably only be given to addresses with different numbers of
6401 registers on machines with lots of registers.
6405 @section Adjusting the Instruction Scheduler
6407 The instruction scheduler may need a fair amount of machine-specific
6408 adjustment in order to produce good code. GCC provides several target
6409 hooks for this purpose. It is usually enough to define just a few of
6410 them: try the first ones in this list first.
6412 @hook TARGET_SCHED_ISSUE_RATE
6413 This hook returns the maximum number of instructions that can ever
6414 issue at the same time on the target machine. The default is one.
6415 Although the insn scheduler can define itself the possibility of issue
6416 an insn on the same cycle, the value can serve as an additional
6417 constraint to issue insns on the same simulated processor cycle (see
6418 hooks @samp{TARGET_SCHED_REORDER} and @samp{TARGET_SCHED_REORDER2}).
6419 This value must be constant over the entire compilation. If you need
6420 it to vary depending on what the instructions are, you must use
6421 @samp{TARGET_SCHED_VARIABLE_ISSUE}.
6424 @hook TARGET_SCHED_VARIABLE_ISSUE
6425 This hook is executed by the scheduler after it has scheduled an insn
6426 from the ready list. It should return the number of insns which can
6427 still be issued in the current cycle. The default is
6428 @samp{@w{@var{more} - 1}} for insns other than @code{CLOBBER} and
6429 @code{USE}, which normally are not counted against the issue rate.
6430 You should define this hook if some insns take more machine resources
6431 than others, so that fewer insns can follow them in the same cycle.
6432 @var{file} is either a null pointer, or a stdio stream to write any
6433 debug output to. @var{verbose} is the verbose level provided by
6434 @option{-fsched-verbose-@var{n}}. @var{insn} is the instruction that
6438 @hook TARGET_SCHED_ADJUST_COST
6439 This function corrects the value of @var{cost} based on the
6440 relationship between @var{insn} and @var{dep_insn} through the
6441 dependence @var{link}. It should return the new value. The default
6442 is to make no adjustment to @var{cost}. This can be used for example
6443 to specify to the scheduler using the traditional pipeline description
6444 that an output- or anti-dependence does not incur the same cost as a
6445 data-dependence. If the scheduler using the automaton based pipeline
6446 description, the cost of anti-dependence is zero and the cost of
6447 output-dependence is maximum of one and the difference of latency
6448 times of the first and the second insns. If these values are not
6449 acceptable, you could use the hook to modify them too. See also
6450 @pxref{Processor pipeline description}.
6453 @hook TARGET_SCHED_ADJUST_PRIORITY
6454 This hook adjusts the integer scheduling priority @var{priority} of
6455 @var{insn}. It should return the new priority. Increase the priority to
6456 execute @var{insn} earlier, reduce the priority to execute @var{insn}
6457 later. Do not define this hook if you do not need to adjust the
6458 scheduling priorities of insns.
6461 @hook TARGET_SCHED_REORDER
6462 This hook is executed by the scheduler after it has scheduled the ready
6463 list, to allow the machine description to reorder it (for example to
6464 combine two small instructions together on @samp{VLIW} machines).
6465 @var{file} is either a null pointer, or a stdio stream to write any
6466 debug output to. @var{verbose} is the verbose level provided by
6467 @option{-fsched-verbose-@var{n}}. @var{ready} is a pointer to the ready
6468 list of instructions that are ready to be scheduled. @var{n_readyp} is
6469 a pointer to the number of elements in the ready list. The scheduler
6470 reads the ready list in reverse order, starting with
6471 @var{ready}[@var{*n_readyp} @minus{} 1] and going to @var{ready}[0]. @var{clock}
6472 is the timer tick of the scheduler. You may modify the ready list and
6473 the number of ready insns. The return value is the number of insns that
6474 can issue this cycle; normally this is just @code{issue_rate}. See also
6475 @samp{TARGET_SCHED_REORDER2}.
6478 @hook TARGET_SCHED_REORDER2
6479 Like @samp{TARGET_SCHED_REORDER}, but called at a different time. That
6480 function is called whenever the scheduler starts a new cycle. This one
6481 is called once per iteration over a cycle, immediately after
6482 @samp{TARGET_SCHED_VARIABLE_ISSUE}; it can reorder the ready list and
6483 return the number of insns to be scheduled in the same cycle. Defining
6484 this hook can be useful if there are frequent situations where
6485 scheduling one insn causes other insns to become ready in the same
6486 cycle. These other insns can then be taken into account properly.
6489 @hook TARGET_SCHED_DEPENDENCIES_EVALUATION_HOOK
6490 This hook is called after evaluation forward dependencies of insns in
6491 chain given by two parameter values (@var{head} and @var{tail}
6492 correspondingly) but before insns scheduling of the insn chain. For
6493 example, it can be used for better insn classification if it requires
6494 analysis of dependencies. This hook can use backward and forward
6495 dependencies of the insn scheduler because they are already
6499 @hook TARGET_SCHED_INIT
6500 This hook is executed by the scheduler at the beginning of each block of
6501 instructions that are to be scheduled. @var{file} is either a null
6502 pointer, or a stdio stream to write any debug output to. @var{verbose}
6503 is the verbose level provided by @option{-fsched-verbose-@var{n}}.
6504 @var{max_ready} is the maximum number of insns in the current scheduling
6505 region that can be live at the same time. This can be used to allocate
6506 scratch space if it is needed, e.g.@: by @samp{TARGET_SCHED_REORDER}.
6509 @hook TARGET_SCHED_FINISH
6510 This hook is executed by the scheduler at the end of each block of
6511 instructions that are to be scheduled. It can be used to perform
6512 cleanup of any actions done by the other scheduling hooks. @var{file}
6513 is either a null pointer, or a stdio stream to write any debug output
6514 to. @var{verbose} is the verbose level provided by
6515 @option{-fsched-verbose-@var{n}}.
6518 @hook TARGET_SCHED_INIT_GLOBAL
6519 This hook is executed by the scheduler after function level initializations.
6520 @var{file} is either a null pointer, or a stdio stream to write any debug output to.
6521 @var{verbose} is the verbose level provided by @option{-fsched-verbose-@var{n}}.
6522 @var{old_max_uid} is the maximum insn uid when scheduling begins.
6525 @hook TARGET_SCHED_FINISH_GLOBAL
6526 This is the cleanup hook corresponding to @code{TARGET_SCHED_INIT_GLOBAL}.
6527 @var{file} is either a null pointer, or a stdio stream to write any debug output to.
6528 @var{verbose} is the verbose level provided by @option{-fsched-verbose-@var{n}}.
6531 @hook TARGET_SCHED_DFA_PRE_CYCLE_INSN
6532 The hook returns an RTL insn. The automaton state used in the
6533 pipeline hazard recognizer is changed as if the insn were scheduled
6534 when the new simulated processor cycle starts. Usage of the hook may
6535 simplify the automaton pipeline description for some @acronym{VLIW}
6536 processors. If the hook is defined, it is used only for the automaton
6537 based pipeline description. The default is not to change the state
6538 when the new simulated processor cycle starts.
6541 @hook TARGET_SCHED_INIT_DFA_PRE_CYCLE_INSN
6542 The hook can be used to initialize data used by the previous hook.
6545 @hook TARGET_SCHED_DFA_POST_CYCLE_INSN
6546 The hook is analogous to @samp{TARGET_SCHED_DFA_PRE_CYCLE_INSN} but used
6547 to changed the state as if the insn were scheduled when the new
6548 simulated processor cycle finishes.
6551 @hook TARGET_SCHED_INIT_DFA_POST_CYCLE_INSN
6552 The hook is analogous to @samp{TARGET_SCHED_INIT_DFA_PRE_CYCLE_INSN} but
6553 used to initialize data used by the previous hook.
6556 @hook TARGET_SCHED_DFA_PRE_ADVANCE_CYCLE
6557 The hook to notify target that the current simulated cycle is about to finish.
6558 The hook is analogous to @samp{TARGET_SCHED_DFA_PRE_CYCLE_INSN} but used
6559 to change the state in more complicated situations - e.g., when advancing
6560 state on a single insn is not enough.
6563 @hook TARGET_SCHED_DFA_POST_ADVANCE_CYCLE
6564 The hook to notify target that new simulated cycle has just started.
6565 The hook is analogous to @samp{TARGET_SCHED_DFA_POST_CYCLE_INSN} but used
6566 to change the state in more complicated situations - e.g., when advancing
6567 state on a single insn is not enough.
6570 @hook TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD
6571 This hook controls better choosing an insn from the ready insn queue
6572 for the @acronym{DFA}-based insn scheduler. Usually the scheduler
6573 chooses the first insn from the queue. If the hook returns a positive
6574 value, an additional scheduler code tries all permutations of
6575 @samp{TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD ()}
6576 subsequent ready insns to choose an insn whose issue will result in
6577 maximal number of issued insns on the same cycle. For the
6578 @acronym{VLIW} processor, the code could actually solve the problem of
6579 packing simple insns into the @acronym{VLIW} insn. Of course, if the
6580 rules of @acronym{VLIW} packing are described in the automaton.
6582 This code also could be used for superscalar @acronym{RISC}
6583 processors. Let us consider a superscalar @acronym{RISC} processor
6584 with 3 pipelines. Some insns can be executed in pipelines @var{A} or
6585 @var{B}, some insns can be executed only in pipelines @var{B} or
6586 @var{C}, and one insn can be executed in pipeline @var{B}. The
6587 processor may issue the 1st insn into @var{A} and the 2nd one into
6588 @var{B}. In this case, the 3rd insn will wait for freeing @var{B}
6589 until the next cycle. If the scheduler issues the 3rd insn the first,
6590 the processor could issue all 3 insns per cycle.
6592 Actually this code demonstrates advantages of the automaton based
6593 pipeline hazard recognizer. We try quickly and easy many insn
6594 schedules to choose the best one.
6596 The default is no multipass scheduling.
6599 @hook TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD
6601 This hook controls what insns from the ready insn queue will be
6602 considered for the multipass insn scheduling. If the hook returns
6603 zero for @var{insn}, the insn will be not chosen to
6606 The default is that any ready insns can be chosen to be issued.
6609 @hook TARGET_SCHED_FIRST_CYCLE_MULTIPASS_BEGIN
6610 This hook prepares the target backend for a new round of multipass
6614 @hook TARGET_SCHED_FIRST_CYCLE_MULTIPASS_ISSUE
6615 This hook is called when multipass scheduling evaluates instruction INSN.
6618 @hook TARGET_SCHED_FIRST_CYCLE_MULTIPASS_BACKTRACK
6619 This is called when multipass scheduling backtracks from evaluation of
6623 @hook TARGET_SCHED_FIRST_CYCLE_MULTIPASS_END
6624 This hook notifies the target about the result of the concluded current
6625 round of multipass scheduling.
6628 @hook TARGET_SCHED_FIRST_CYCLE_MULTIPASS_INIT
6629 This hook initializes target-specific data used in multipass scheduling.
6632 @hook TARGET_SCHED_FIRST_CYCLE_MULTIPASS_FINI
6633 This hook finalizes target-specific data used in multipass scheduling.
6636 @hook TARGET_SCHED_DFA_NEW_CYCLE
6637 This hook is called by the insn scheduler before issuing @var{insn}
6638 on cycle @var{clock}. If the hook returns nonzero,
6639 @var{insn} is not issued on this processor cycle. Instead,
6640 the processor cycle is advanced. If *@var{sort_p}
6641 is zero, the insn ready queue is not sorted on the new cycle
6642 start as usually. @var{dump} and @var{verbose} specify the file and
6643 verbosity level to use for debugging output.
6644 @var{last_clock} and @var{clock} are, respectively, the
6645 processor cycle on which the previous insn has been issued,
6646 and the current processor cycle.
6649 @hook TARGET_SCHED_IS_COSTLY_DEPENDENCE
6650 This hook is used to define which dependences are considered costly by
6651 the target, so costly that it is not advisable to schedule the insns that
6652 are involved in the dependence too close to one another. The parameters
6653 to this hook are as follows: The first parameter @var{_dep} is the dependence
6654 being evaluated. The second parameter @var{cost} is the cost of the
6655 dependence as estimated by the scheduler, and the third
6656 parameter @var{distance} is the distance in cycles between the two insns.
6657 The hook returns @code{true} if considering the distance between the two
6658 insns the dependence between them is considered costly by the target,
6659 and @code{false} otherwise.
6661 Defining this hook can be useful in multiple-issue out-of-order machines,
6662 where (a) it's practically hopeless to predict the actual data/resource
6663 delays, however: (b) there's a better chance to predict the actual grouping
6664 that will be formed, and (c) correctly emulating the grouping can be very
6665 important. In such targets one may want to allow issuing dependent insns
6666 closer to one another---i.e., closer than the dependence distance; however,
6667 not in cases of ``costly dependences'', which this hooks allows to define.
6670 @hook TARGET_SCHED_H_I_D_EXTENDED
6671 This hook is called by the insn scheduler after emitting a new instruction to
6672 the instruction stream. The hook notifies a target backend to extend its
6673 per instruction data structures.
6676 @hook TARGET_SCHED_ALLOC_SCHED_CONTEXT
6677 Return a pointer to a store large enough to hold target scheduling context.
6680 @hook TARGET_SCHED_INIT_SCHED_CONTEXT
6681 Initialize store pointed to by @var{tc} to hold target scheduling context.
6682 It @var{clean_p} is true then initialize @var{tc} as if scheduler is at the
6683 beginning of the block. Otherwise, copy the current context into @var{tc}.
6686 @hook TARGET_SCHED_SET_SCHED_CONTEXT
6687 Copy target scheduling context pointed to by @var{tc} to the current context.
6690 @hook TARGET_SCHED_CLEAR_SCHED_CONTEXT
6691 Deallocate internal data in target scheduling context pointed to by @var{tc}.
6694 @hook TARGET_SCHED_FREE_SCHED_CONTEXT
6695 Deallocate a store for target scheduling context pointed to by @var{tc}.
6698 @hook TARGET_SCHED_SPECULATE_INSN
6699 This hook is called by the insn scheduler when @var{insn} has only
6700 speculative dependencies and therefore can be scheduled speculatively.
6701 The hook is used to check if the pattern of @var{insn} has a speculative
6702 version and, in case of successful check, to generate that speculative
6703 pattern. The hook should return 1, if the instruction has a speculative form,
6704 or @minus{}1, if it doesn't. @var{request} describes the type of requested
6705 speculation. If the return value equals 1 then @var{new_pat} is assigned
6706 the generated speculative pattern.
6709 @hook TARGET_SCHED_NEEDS_BLOCK_P
6710 This hook is called by the insn scheduler during generation of recovery code
6711 for @var{insn}. It should return @code{true}, if the corresponding check
6712 instruction should branch to recovery code, or @code{false} otherwise.
6715 @hook TARGET_SCHED_GEN_SPEC_CHECK
6716 This hook is called by the insn scheduler to generate a pattern for recovery
6717 check instruction. If @var{mutate_p} is zero, then @var{insn} is a
6718 speculative instruction for which the check should be generated.
6719 @var{label} is either a label of a basic block, where recovery code should
6720 be emitted, or a null pointer, when requested check doesn't branch to
6721 recovery code (a simple check). If @var{mutate_p} is nonzero, then
6722 a pattern for a branchy check corresponding to a simple check denoted by
6723 @var{insn} should be generated. In this case @var{label} can't be null.
6726 @hook TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC
6727 This hook is used as a workaround for
6728 @samp{TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD} not being
6729 called on the first instruction of the ready list. The hook is used to
6730 discard speculative instructions that stand first in the ready list from
6731 being scheduled on the current cycle. If the hook returns @code{false},
6732 @var{insn} will not be chosen to be issued.
6733 For non-speculative instructions,
6734 the hook should always return @code{true}. For example, in the ia64 backend
6735 the hook is used to cancel data speculative insns when the ALAT table
6739 @hook TARGET_SCHED_SET_SCHED_FLAGS
6740 This hook is used by the insn scheduler to find out what features should be
6742 The structure *@var{spec_info} should be filled in by the target.
6743 The structure describes speculation types that can be used in the scheduler.
6746 @hook TARGET_SCHED_SMS_RES_MII
6747 This hook is called by the swing modulo scheduler to calculate a
6748 resource-based lower bound which is based on the resources available in
6749 the machine and the resources required by each instruction. The target
6750 backend can use @var{g} to calculate such bound. A very simple lower
6751 bound will be used in case this hook is not implemented: the total number
6752 of instructions divided by the issue rate.
6755 @hook TARGET_SCHED_DISPATCH
6756 This hook is called by Haifa Scheduler. It returns true if dispatch scheduling
6757 is supported in hardware and the condition specified in the parameter is true.
6760 @hook TARGET_SCHED_DISPATCH_DO
6761 This hook is called by Haifa Scheduler. It performs the operation specified
6762 in its second parameter.
6765 @hook TARGET_SCHED_EXPOSED_PIPELINE
6767 @hook TARGET_SCHED_REASSOCIATION_WIDTH
6770 @section Dividing the Output into Sections (Texts, Data, @dots{})
6771 @c the above section title is WAY too long. maybe cut the part between
6772 @c the (...)? --mew 10feb93
6774 An object file is divided into sections containing different types of
6775 data. In the most common case, there are three sections: the @dfn{text
6776 section}, which holds instructions and read-only data; the @dfn{data
6777 section}, which holds initialized writable data; and the @dfn{bss
6778 section}, which holds uninitialized data. Some systems have other kinds
6781 @file{varasm.c} provides several well-known sections, such as
6782 @code{text_section}, @code{data_section} and @code{bss_section}.
6783 The normal way of controlling a @code{@var{foo}_section} variable
6784 is to define the associated @code{@var{FOO}_SECTION_ASM_OP} macro,
6785 as described below. The macros are only read once, when @file{varasm.c}
6786 initializes itself, so their values must be run-time constants.
6787 They may however depend on command-line flags.
6789 @emph{Note:} Some run-time files, such @file{crtstuff.c}, also make
6790 use of the @code{@var{FOO}_SECTION_ASM_OP} macros, and expect them
6791 to be string literals.
6793 Some assemblers require a different string to be written every time a
6794 section is selected. If your assembler falls into this category, you
6795 should define the @code{TARGET_ASM_INIT_SECTIONS} hook and use
6796 @code{get_unnamed_section} to set up the sections.
6798 You must always create a @code{text_section}, either by defining
6799 @code{TEXT_SECTION_ASM_OP} or by initializing @code{text_section}
6800 in @code{TARGET_ASM_INIT_SECTIONS}. The same is true of
6801 @code{data_section} and @code{DATA_SECTION_ASM_OP}. If you do not
6802 create a distinct @code{readonly_data_section}, the default is to
6803 reuse @code{text_section}.
6805 All the other @file{varasm.c} sections are optional, and are null
6806 if the target does not provide them.
6808 @defmac TEXT_SECTION_ASM_OP
6809 A C expression whose value is a string, including spacing, containing the
6810 assembler operation that should precede instructions and read-only data.
6811 Normally @code{"\t.text"} is right.
6814 @defmac HOT_TEXT_SECTION_NAME
6815 If defined, a C string constant for the name of the section containing most
6816 frequently executed functions of the program. If not defined, GCC will provide
6817 a default definition if the target supports named sections.
6820 @defmac UNLIKELY_EXECUTED_TEXT_SECTION_NAME
6821 If defined, a C string constant for the name of the section containing unlikely
6822 executed functions in the program.
6825 @defmac DATA_SECTION_ASM_OP
6826 A C expression whose value is a string, including spacing, containing the
6827 assembler operation to identify the following data as writable initialized
6828 data. Normally @code{"\t.data"} is right.
6831 @defmac SDATA_SECTION_ASM_OP
6832 If defined, a C expression whose value is a string, including spacing,
6833 containing the assembler operation to identify the following data as
6834 initialized, writable small data.
6837 @defmac READONLY_DATA_SECTION_ASM_OP
6838 A C expression whose value is a string, including spacing, containing the
6839 assembler operation to identify the following data as read-only initialized
6843 @defmac BSS_SECTION_ASM_OP
6844 If defined, a C expression whose value is a string, including spacing,
6845 containing the assembler operation to identify the following data as
6846 uninitialized global data. If not defined, and
6847 @code{ASM_OUTPUT_ALIGNED_BSS} not defined,
6848 uninitialized global data will be output in the data section if
6849 @option{-fno-common} is passed, otherwise @code{ASM_OUTPUT_COMMON} will be
6853 @defmac SBSS_SECTION_ASM_OP
6854 If defined, a C expression whose value is a string, including spacing,
6855 containing the assembler operation to identify the following data as
6856 uninitialized, writable small data.
6859 @defmac TLS_COMMON_ASM_OP
6860 If defined, a C expression whose value is a string containing the
6861 assembler operation to identify the following data as thread-local
6862 common data. The default is @code{".tls_common"}.
6865 @defmac TLS_SECTION_ASM_FLAG
6866 If defined, a C expression whose value is a character constant
6867 containing the flag used to mark a section as a TLS section. The
6868 default is @code{'T'}.
6871 @defmac INIT_SECTION_ASM_OP
6872 If defined, a C expression whose value is a string, including spacing,
6873 containing the assembler operation to identify the following data as
6874 initialization code. If not defined, GCC will assume such a section does
6875 not exist. This section has no corresponding @code{init_section}
6876 variable; it is used entirely in runtime code.
6879 @defmac FINI_SECTION_ASM_OP
6880 If defined, a C expression whose value is a string, including spacing,
6881 containing the assembler operation to identify the following data as
6882 finalization code. If not defined, GCC will assume such a section does
6883 not exist. This section has no corresponding @code{fini_section}
6884 variable; it is used entirely in runtime code.
6887 @defmac INIT_ARRAY_SECTION_ASM_OP
6888 If defined, a C expression whose value is a string, including spacing,
6889 containing the assembler operation to identify the following data as
6890 part of the @code{.init_array} (or equivalent) section. If not
6891 defined, GCC will assume such a section does not exist. Do not define
6892 both this macro and @code{INIT_SECTION_ASM_OP}.
6895 @defmac FINI_ARRAY_SECTION_ASM_OP
6896 If defined, a C expression whose value is a string, including spacing,
6897 containing the assembler operation to identify the following data as
6898 part of the @code{.fini_array} (or equivalent) section. If not
6899 defined, GCC will assume such a section does not exist. Do not define
6900 both this macro and @code{FINI_SECTION_ASM_OP}.
6903 @defmac CRT_CALL_STATIC_FUNCTION (@var{section_op}, @var{function})
6904 If defined, an ASM statement that switches to a different section
6905 via @var{section_op}, calls @var{function}, and switches back to
6906 the text section. This is used in @file{crtstuff.c} if
6907 @code{INIT_SECTION_ASM_OP} or @code{FINI_SECTION_ASM_OP} to calls
6908 to initialization and finalization functions from the init and fini
6909 sections. By default, this macro uses a simple function call. Some
6910 ports need hand-crafted assembly code to avoid dependencies on
6911 registers initialized in the function prologue or to ensure that
6912 constant pools don't end up too far way in the text section.
6915 @defmac TARGET_LIBGCC_SDATA_SECTION
6916 If defined, a string which names the section into which small
6917 variables defined in crtstuff and libgcc should go. This is useful
6918 when the target has options for optimizing access to small data, and
6919 you want the crtstuff and libgcc routines to be conservative in what
6920 they expect of your application yet liberal in what your application
6921 expects. For example, for targets with a @code{.sdata} section (like
6922 MIPS), you could compile crtstuff with @code{-G 0} so that it doesn't
6923 require small data support from your application, but use this macro
6924 to put small data into @code{.sdata} so that your application can
6925 access these variables whether it uses small data or not.
6928 @defmac FORCE_CODE_SECTION_ALIGN
6929 If defined, an ASM statement that aligns a code section to some
6930 arbitrary boundary. This is used to force all fragments of the
6931 @code{.init} and @code{.fini} sections to have to same alignment
6932 and thus prevent the linker from having to add any padding.
6935 @defmac JUMP_TABLES_IN_TEXT_SECTION
6936 Define this macro to be an expression with a nonzero value if jump
6937 tables (for @code{tablejump} insns) should be output in the text
6938 section, along with the assembler instructions. Otherwise, the
6939 readonly data section is used.
6941 This macro is irrelevant if there is no separate readonly data section.
6944 @hook TARGET_ASM_INIT_SECTIONS
6945 Define this hook if you need to do something special to set up the
6946 @file{varasm.c} sections, or if your target has some special sections
6947 of its own that you need to create.
6949 GCC calls this hook after processing the command line, but before writing
6950 any assembly code, and before calling any of the section-returning hooks
6954 @hook TARGET_ASM_RELOC_RW_MASK
6955 Return a mask describing how relocations should be treated when
6956 selecting sections. Bit 1 should be set if global relocations
6957 should be placed in a read-write section; bit 0 should be set if
6958 local relocations should be placed in a read-write section.
6960 The default version of this function returns 3 when @option{-fpic}
6961 is in effect, and 0 otherwise. The hook is typically redefined
6962 when the target cannot support (some kinds of) dynamic relocations
6963 in read-only sections even in executables.
6966 @hook TARGET_ASM_SELECT_SECTION
6967 Return the section into which @var{exp} should be placed. You can
6968 assume that @var{exp} is either a @code{VAR_DECL} node or a constant of
6969 some sort. @var{reloc} indicates whether the initial value of @var{exp}
6970 requires link-time relocations. Bit 0 is set when variable contains
6971 local relocations only, while bit 1 is set for global relocations.
6972 @var{align} is the constant alignment in bits.
6974 The default version of this function takes care of putting read-only
6975 variables in @code{readonly_data_section}.
6977 See also @var{USE_SELECT_SECTION_FOR_FUNCTIONS}.
6980 @defmac USE_SELECT_SECTION_FOR_FUNCTIONS
6981 Define this macro if you wish TARGET_ASM_SELECT_SECTION to be called
6982 for @code{FUNCTION_DECL}s as well as for variables and constants.
6984 In the case of a @code{FUNCTION_DECL}, @var{reloc} will be zero if the
6985 function has been determined to be likely to be called, and nonzero if
6986 it is unlikely to be called.
6989 @hook TARGET_ASM_UNIQUE_SECTION
6990 Build up a unique section name, expressed as a @code{STRING_CST} node,
6991 and assign it to @samp{DECL_SECTION_NAME (@var{decl})}.
6992 As with @code{TARGET_ASM_SELECT_SECTION}, @var{reloc} indicates whether
6993 the initial value of @var{exp} requires link-time relocations.
6995 The default version of this function appends the symbol name to the
6996 ELF section name that would normally be used for the symbol. For
6997 example, the function @code{foo} would be placed in @code{.text.foo}.
6998 Whatever the actual target object format, this is often good enough.
7001 @hook TARGET_ASM_FUNCTION_RODATA_SECTION
7002 Return the readonly data section associated with
7003 @samp{DECL_SECTION_NAME (@var{decl})}.
7004 The default version of this function selects @code{.gnu.linkonce.r.name} if
7005 the function's section is @code{.gnu.linkonce.t.name}, @code{.rodata.name}
7006 if function is in @code{.text.name}, and the normal readonly-data section
7010 @hook TARGET_ASM_MERGEABLE_RODATA_PREFIX
7012 @hook TARGET_ASM_SELECT_RTX_SECTION
7013 Return the section into which a constant @var{x}, of mode @var{mode},
7014 should be placed. You can assume that @var{x} is some kind of
7015 constant in RTL@. The argument @var{mode} is redundant except in the
7016 case of a @code{const_int} rtx. @var{align} is the constant alignment
7019 The default version of this function takes care of putting symbolic
7020 constants in @code{flag_pic} mode in @code{data_section} and everything
7021 else in @code{readonly_data_section}.
7024 @hook TARGET_MANGLE_DECL_ASSEMBLER_NAME
7025 Define this hook if you need to postprocess the assembler name generated
7026 by target-independent code. The @var{id} provided to this hook will be
7027 the computed name (e.g., the macro @code{DECL_NAME} of the @var{decl} in C,
7028 or the mangled name of the @var{decl} in C++). The return value of the
7029 hook is an @code{IDENTIFIER_NODE} for the appropriate mangled name on
7030 your target system. The default implementation of this hook just
7031 returns the @var{id} provided.
7034 @hook TARGET_ENCODE_SECTION_INFO
7035 Define this hook if references to a symbol or a constant must be
7036 treated differently depending on something about the variable or
7037 function named by the symbol (such as what section it is in).
7039 The hook is executed immediately after rtl has been created for
7040 @var{decl}, which may be a variable or function declaration or
7041 an entry in the constant pool. In either case, @var{rtl} is the
7042 rtl in question. Do @emph{not} use @code{DECL_RTL (@var{decl})}
7043 in this hook; that field may not have been initialized yet.
7045 In the case of a constant, it is safe to assume that the rtl is
7046 a @code{mem} whose address is a @code{symbol_ref}. Most decls
7047 will also have this form, but that is not guaranteed. Global
7048 register variables, for instance, will have a @code{reg} for their
7049 rtl. (Normally the right thing to do with such unusual rtl is
7052 The @var{new_decl_p} argument will be true if this is the first time
7053 that @code{TARGET_ENCODE_SECTION_INFO} has been invoked on this decl. It will
7054 be false for subsequent invocations, which will happen for duplicate
7055 declarations. Whether or not anything must be done for the duplicate
7056 declaration depends on whether the hook examines @code{DECL_ATTRIBUTES}.
7057 @var{new_decl_p} is always true when the hook is called for a constant.
7059 @cindex @code{SYMBOL_REF_FLAG}, in @code{TARGET_ENCODE_SECTION_INFO}
7060 The usual thing for this hook to do is to record flags in the
7061 @code{symbol_ref}, using @code{SYMBOL_REF_FLAG} or @code{SYMBOL_REF_FLAGS}.
7062 Historically, the name string was modified if it was necessary to
7063 encode more than one bit of information, but this practice is now
7064 discouraged; use @code{SYMBOL_REF_FLAGS}.
7066 The default definition of this hook, @code{default_encode_section_info}
7067 in @file{varasm.c}, sets a number of commonly-useful bits in
7068 @code{SYMBOL_REF_FLAGS}. Check whether the default does what you need
7069 before overriding it.
7072 @hook TARGET_STRIP_NAME_ENCODING
7073 Decode @var{name} and return the real name part, sans
7074 the characters that @code{TARGET_ENCODE_SECTION_INFO}
7078 @hook TARGET_IN_SMALL_DATA_P
7079 Returns true if @var{exp} should be placed into a ``small data'' section.
7080 The default version of this hook always returns false.
7083 @hook TARGET_HAVE_SRODATA_SECTION
7084 Contains the value true if the target places read-only
7085 ``small data'' into a separate section. The default value is false.
7088 @hook TARGET_PROFILE_BEFORE_PROLOGUE
7090 @hook TARGET_BINDS_LOCAL_P
7091 Returns true if @var{exp} names an object for which name resolution
7092 rules must resolve to the current ``module'' (dynamic shared library
7093 or executable image).
7095 The default version of this hook implements the name resolution rules
7096 for ELF, which has a looser model of global name binding than other
7097 currently supported object file formats.
7100 @hook TARGET_HAVE_TLS
7101 Contains the value true if the target supports thread-local storage.
7102 The default value is false.
7107 @section Position Independent Code
7108 @cindex position independent code
7111 This section describes macros that help implement generation of position
7112 independent code. Simply defining these macros is not enough to
7113 generate valid PIC; you must also add support to the hook
7114 @code{TARGET_LEGITIMATE_ADDRESS_P} and to the macro
7115 @code{PRINT_OPERAND_ADDRESS}, as well as @code{LEGITIMIZE_ADDRESS}. You
7116 must modify the definition of @samp{movsi} to do something appropriate
7117 when the source operand contains a symbolic address. You may also
7118 need to alter the handling of switch statements so that they use
7120 @c i rearranged the order of the macros above to try to force one of
7121 @c them to the next line, to eliminate an overfull hbox. --mew 10feb93
7123 @defmac PIC_OFFSET_TABLE_REGNUM
7124 The register number of the register used to address a table of static
7125 data addresses in memory. In some cases this register is defined by a
7126 processor's ``application binary interface'' (ABI)@. When this macro
7127 is defined, RTL is generated for this register once, as with the stack
7128 pointer and frame pointer registers. If this macro is not defined, it
7129 is up to the machine-dependent files to allocate such a register (if
7130 necessary). Note that this register must be fixed when in use (e.g.@:
7131 when @code{flag_pic} is true).
7134 @defmac PIC_OFFSET_TABLE_REG_CALL_CLOBBERED
7135 A C expression that is nonzero if the register defined by
7136 @code{PIC_OFFSET_TABLE_REGNUM} is clobbered by calls. If not defined,
7137 the default is zero. Do not define
7138 this macro if @code{PIC_OFFSET_TABLE_REGNUM} is not defined.
7141 @defmac LEGITIMATE_PIC_OPERAND_P (@var{x})
7142 A C expression that is nonzero if @var{x} is a legitimate immediate
7143 operand on the target machine when generating position independent code.
7144 You can assume that @var{x} satisfies @code{CONSTANT_P}, so you need not
7145 check this. You can also assume @var{flag_pic} is true, so you need not
7146 check it either. You need not define this macro if all constants
7147 (including @code{SYMBOL_REF}) can be immediate operands when generating
7148 position independent code.
7151 @node Assembler Format
7152 @section Defining the Output Assembler Language
7154 This section describes macros whose principal purpose is to describe how
7155 to write instructions in assembler language---rather than what the
7159 * File Framework:: Structural information for the assembler file.
7160 * Data Output:: Output of constants (numbers, strings, addresses).
7161 * Uninitialized Data:: Output of uninitialized variables.
7162 * Label Output:: Output and generation of labels.
7163 * Initialization:: General principles of initialization
7164 and termination routines.
7165 * Macros for Initialization::
7166 Specific macros that control the handling of
7167 initialization and termination routines.
7168 * Instruction Output:: Output of actual instructions.
7169 * Dispatch Tables:: Output of jump tables.
7170 * Exception Region Output:: Output of exception region code.
7171 * Alignment Output:: Pseudo ops for alignment and skipping data.
7174 @node File Framework
7175 @subsection The Overall Framework of an Assembler File
7176 @cindex assembler format
7177 @cindex output of assembler code
7179 @c prevent bad page break with this line
7180 This describes the overall framework of an assembly file.
7182 @findex default_file_start
7183 @hook TARGET_ASM_FILE_START
7184 Output to @code{asm_out_file} any text which the assembler expects to
7185 find at the beginning of a file. The default behavior is controlled
7186 by two flags, documented below. Unless your target's assembler is
7187 quite unusual, if you override the default, you should call
7188 @code{default_file_start} at some point in your target hook. This
7189 lets other target files rely on these variables.
7192 @hook TARGET_ASM_FILE_START_APP_OFF
7193 If this flag is true, the text of the macro @code{ASM_APP_OFF} will be
7194 printed as the very first line in the assembly file, unless
7195 @option{-fverbose-asm} is in effect. (If that macro has been defined
7196 to the empty string, this variable has no effect.) With the normal
7197 definition of @code{ASM_APP_OFF}, the effect is to notify the GNU
7198 assembler that it need not bother stripping comments or extra
7199 whitespace from its input. This allows it to work a bit faster.
7201 The default is false. You should not set it to true unless you have
7202 verified that your port does not generate any extra whitespace or
7203 comments that will cause GAS to issue errors in NO_APP mode.
7206 @hook TARGET_ASM_FILE_START_FILE_DIRECTIVE
7207 If this flag is true, @code{output_file_directive} will be called
7208 for the primary source file, immediately after printing
7209 @code{ASM_APP_OFF} (if that is enabled). Most ELF assemblers expect
7210 this to be done. The default is false.
7213 @hook TARGET_ASM_FILE_END
7214 Output to @code{asm_out_file} any text which the assembler expects
7215 to find at the end of a file. The default is to output nothing.
7218 @deftypefun void file_end_indicate_exec_stack ()
7219 Some systems use a common convention, the @samp{.note.GNU-stack}
7220 special section, to indicate whether or not an object file relies on
7221 the stack being executable. If your system uses this convention, you
7222 should define @code{TARGET_ASM_FILE_END} to this function. If you
7223 need to do other things in that hook, have your hook function call
7227 @hook TARGET_ASM_LTO_START
7228 Output to @code{asm_out_file} any text which the assembler expects
7229 to find at the start of an LTO section. The default is to output
7233 @hook TARGET_ASM_LTO_END
7234 Output to @code{asm_out_file} any text which the assembler expects
7235 to find at the end of an LTO section. The default is to output
7239 @hook TARGET_ASM_CODE_END
7240 Output to @code{asm_out_file} any text which is needed before emitting
7241 unwind info and debug info at the end of a file. Some targets emit
7242 here PIC setup thunks that cannot be emitted at the end of file,
7243 because they couldn't have unwind info then. The default is to output
7247 @defmac ASM_COMMENT_START
7248 A C string constant describing how to begin a comment in the target
7249 assembler language. The compiler assumes that the comment will end at
7250 the end of the line.
7254 A C string constant for text to be output before each @code{asm}
7255 statement or group of consecutive ones. Normally this is
7256 @code{"#APP"}, which is a comment that has no effect on most
7257 assemblers but tells the GNU assembler that it must check the lines
7258 that follow for all valid assembler constructs.
7262 A C string constant for text to be output after each @code{asm}
7263 statement or group of consecutive ones. Normally this is
7264 @code{"#NO_APP"}, which tells the GNU assembler to resume making the
7265 time-saving assumptions that are valid for ordinary compiler output.
7268 @defmac ASM_OUTPUT_SOURCE_FILENAME (@var{stream}, @var{name})
7269 A C statement to output COFF information or DWARF debugging information
7270 which indicates that filename @var{name} is the current source file to
7271 the stdio stream @var{stream}.
7273 This macro need not be defined if the standard form of output
7274 for the file format in use is appropriate.
7277 @hook TARGET_ASM_OUTPUT_SOURCE_FILENAME
7279 @defmac OUTPUT_QUOTED_STRING (@var{stream}, @var{string})
7280 A C statement to output the string @var{string} to the stdio stream
7281 @var{stream}. If you do not call the function @code{output_quoted_string}
7282 in your config files, GCC will only call it to output filenames to
7283 the assembler source. So you can use it to canonicalize the format
7284 of the filename using this macro.
7287 @defmac ASM_OUTPUT_IDENT (@var{stream}, @var{string})
7288 A C statement to output something to the assembler file to handle a
7289 @samp{#ident} directive containing the text @var{string}. If this
7290 macro is not defined, nothing is output for a @samp{#ident} directive.
7293 @hook TARGET_ASM_NAMED_SECTION
7294 Output assembly directives to switch to section @var{name}. The section
7295 should have attributes as specified by @var{flags}, which is a bit mask
7296 of the @code{SECTION_*} flags defined in @file{output.h}. If @var{decl}
7297 is non-NULL, it is the @code{VAR_DECL} or @code{FUNCTION_DECL} with which
7298 this section is associated.
7301 @hook TARGET_ASM_FUNCTION_SECTION
7302 Return preferred text (sub)section for function @var{decl}.
7303 Main purpose of this function is to separate cold, normal and hot
7304 functions. @var{startup} is true when function is known to be used only
7305 at startup (from static constructors or it is @code{main()}).
7306 @var{exit} is true when function is known to be used only at exit
7307 (from static destructors).
7308 Return NULL if function should go to default text section.
7311 @hook TARGET_ASM_FUNCTION_SWITCHED_TEXT_SECTIONS
7313 @hook TARGET_HAVE_NAMED_SECTIONS
7314 This flag is true if the target supports @code{TARGET_ASM_NAMED_SECTION}.
7315 It must not be modified by command-line option processing.
7318 @anchor{TARGET_HAVE_SWITCHABLE_BSS_SECTIONS}
7319 @hook TARGET_HAVE_SWITCHABLE_BSS_SECTIONS
7320 This flag is true if we can create zeroed data by switching to a BSS
7321 section and then using @code{ASM_OUTPUT_SKIP} to allocate the space.
7322 This is true on most ELF targets.
7325 @hook TARGET_SECTION_TYPE_FLAGS
7326 Choose a set of section attributes for use by @code{TARGET_ASM_NAMED_SECTION}
7327 based on a variable or function decl, a section name, and whether or not the
7328 declaration's initializer may contain runtime relocations. @var{decl} may be
7329 null, in which case read-write data should be assumed.
7331 The default version of this function handles choosing code vs data,
7332 read-only vs read-write data, and @code{flag_pic}. You should only
7333 need to override this if your target has special flags that might be
7334 set via @code{__attribute__}.
7337 @hook TARGET_ASM_RECORD_GCC_SWITCHES
7338 Provides the target with the ability to record the gcc command line
7339 switches that have been passed to the compiler, and options that are
7340 enabled. The @var{type} argument specifies what is being recorded.
7341 It can take the following values:
7344 @item SWITCH_TYPE_PASSED
7345 @var{text} is a command line switch that has been set by the user.
7347 @item SWITCH_TYPE_ENABLED
7348 @var{text} is an option which has been enabled. This might be as a
7349 direct result of a command line switch, or because it is enabled by
7350 default or because it has been enabled as a side effect of a different
7351 command line switch. For example, the @option{-O2} switch enables
7352 various different individual optimization passes.
7354 @item SWITCH_TYPE_DESCRIPTIVE
7355 @var{text} is either NULL or some descriptive text which should be
7356 ignored. If @var{text} is NULL then it is being used to warn the
7357 target hook that either recording is starting or ending. The first
7358 time @var{type} is SWITCH_TYPE_DESCRIPTIVE and @var{text} is NULL, the
7359 warning is for start up and the second time the warning is for
7360 wind down. This feature is to allow the target hook to make any
7361 necessary preparations before it starts to record switches and to
7362 perform any necessary tidying up after it has finished recording
7365 @item SWITCH_TYPE_LINE_START
7366 This option can be ignored by this target hook.
7368 @item SWITCH_TYPE_LINE_END
7369 This option can be ignored by this target hook.
7372 The hook's return value must be zero. Other return values may be
7373 supported in the future.
7375 By default this hook is set to NULL, but an example implementation is
7376 provided for ELF based targets. Called @var{elf_record_gcc_switches},
7377 it records the switches as ASCII text inside a new, string mergeable
7378 section in the assembler output file. The name of the new section is
7379 provided by the @code{TARGET_ASM_RECORD_GCC_SWITCHES_SECTION} target
7383 @hook TARGET_ASM_RECORD_GCC_SWITCHES_SECTION
7384 This is the name of the section that will be created by the example
7385 ELF implementation of the @code{TARGET_ASM_RECORD_GCC_SWITCHES} target
7391 @subsection Output of Data
7394 @hook TARGET_ASM_BYTE_OP
7395 @deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_HI_OP
7396 @deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_SI_OP
7397 @deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_DI_OP
7398 @deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_TI_OP
7399 @deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_HI_OP
7400 @deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_SI_OP
7401 @deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_DI_OP
7402 @deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_TI_OP
7403 These hooks specify assembly directives for creating certain kinds
7404 of integer object. The @code{TARGET_ASM_BYTE_OP} directive creates a
7405 byte-sized object, the @code{TARGET_ASM_ALIGNED_HI_OP} one creates an
7406 aligned two-byte object, and so on. Any of the hooks may be
7407 @code{NULL}, indicating that no suitable directive is available.
7409 The compiler will print these strings at the start of a new line,
7410 followed immediately by the object's initial value. In most cases,
7411 the string should contain a tab, a pseudo-op, and then another tab.
7414 @hook TARGET_ASM_INTEGER
7415 The @code{assemble_integer} function uses this hook to output an
7416 integer object. @var{x} is the object's value, @var{size} is its size
7417 in bytes and @var{aligned_p} indicates whether it is aligned. The
7418 function should return @code{true} if it was able to output the
7419 object. If it returns false, @code{assemble_integer} will try to
7420 split the object into smaller parts.
7422 The default implementation of this hook will use the
7423 @code{TARGET_ASM_BYTE_OP} family of strings, returning @code{false}
7424 when the relevant string is @code{NULL}.
7427 @hook TARGET_ASM_OUTPUT_ADDR_CONST_EXTRA
7428 A target hook to recognize @var{rtx} patterns that @code{output_addr_const}
7429 can't deal with, and output assembly code to @var{file} corresponding to
7430 the pattern @var{x}. This may be used to allow machine-dependent
7431 @code{UNSPEC}s to appear within constants.
7433 If target hook fails to recognize a pattern, it must return @code{false},
7434 so that a standard error message is printed. If it prints an error message
7435 itself, by calling, for example, @code{output_operand_lossage}, it may just
7439 @defmac ASM_OUTPUT_ASCII (@var{stream}, @var{ptr}, @var{len})
7440 A C statement to output to the stdio stream @var{stream} an assembler
7441 instruction to assemble a string constant containing the @var{len}
7442 bytes at @var{ptr}. @var{ptr} will be a C expression of type
7443 @code{char *} and @var{len} a C expression of type @code{int}.
7445 If the assembler has a @code{.ascii} pseudo-op as found in the
7446 Berkeley Unix assembler, do not define the macro
7447 @code{ASM_OUTPUT_ASCII}.
7450 @defmac ASM_OUTPUT_FDESC (@var{stream}, @var{decl}, @var{n})
7451 A C statement to output word @var{n} of a function descriptor for
7452 @var{decl}. This must be defined if @code{TARGET_VTABLE_USES_DESCRIPTORS}
7453 is defined, and is otherwise unused.
7456 @defmac CONSTANT_POOL_BEFORE_FUNCTION
7457 You may define this macro as a C expression. You should define the
7458 expression to have a nonzero value if GCC should output the constant
7459 pool for a function before the code for the function, or a zero value if
7460 GCC should output the constant pool after the function. If you do
7461 not define this macro, the usual case, GCC will output the constant
7462 pool before the function.
7465 @defmac ASM_OUTPUT_POOL_PROLOGUE (@var{file}, @var{funname}, @var{fundecl}, @var{size})
7466 A C statement to output assembler commands to define the start of the
7467 constant pool for a function. @var{funname} is a string giving
7468 the name of the function. Should the return type of the function
7469 be required, it can be obtained via @var{fundecl}. @var{size}
7470 is the size, in bytes, of the constant pool that will be written
7471 immediately after this call.
7473 If no constant-pool prefix is required, the usual case, this macro need
7477 @defmac ASM_OUTPUT_SPECIAL_POOL_ENTRY (@var{file}, @var{x}, @var{mode}, @var{align}, @var{labelno}, @var{jumpto})
7478 A C statement (with or without semicolon) to output a constant in the
7479 constant pool, if it needs special treatment. (This macro need not do
7480 anything for RTL expressions that can be output normally.)
7482 The argument @var{file} is the standard I/O stream to output the
7483 assembler code on. @var{x} is the RTL expression for the constant to
7484 output, and @var{mode} is the machine mode (in case @var{x} is a
7485 @samp{const_int}). @var{align} is the required alignment for the value
7486 @var{x}; you should output an assembler directive to force this much
7489 The argument @var{labelno} is a number to use in an internal label for
7490 the address of this pool entry. The definition of this macro is
7491 responsible for outputting the label definition at the proper place.
7492 Here is how to do this:
7495 @code{(*targetm.asm_out.internal_label)} (@var{file}, "LC", @var{labelno});
7498 When you output a pool entry specially, you should end with a
7499 @code{goto} to the label @var{jumpto}. This will prevent the same pool
7500 entry from being output a second time in the usual manner.
7502 You need not define this macro if it would do nothing.
7505 @defmac ASM_OUTPUT_POOL_EPILOGUE (@var{file} @var{funname} @var{fundecl} @var{size})
7506 A C statement to output assembler commands to at the end of the constant
7507 pool for a function. @var{funname} is a string giving the name of the
7508 function. Should the return type of the function be required, you can
7509 obtain it via @var{fundecl}. @var{size} is the size, in bytes, of the
7510 constant pool that GCC wrote immediately before this call.
7512 If no constant-pool epilogue is required, the usual case, you need not
7516 @defmac IS_ASM_LOGICAL_LINE_SEPARATOR (@var{C}, @var{STR})
7517 Define this macro as a C expression which is nonzero if @var{C} is
7518 used as a logical line separator by the assembler. @var{STR} points
7519 to the position in the string where @var{C} was found; this can be used if
7520 a line separator uses multiple characters.
7522 If you do not define this macro, the default is that only
7523 the character @samp{;} is treated as a logical line separator.
7526 @hook TARGET_ASM_OPEN_PAREN
7527 These target hooks are C string constants, describing the syntax in the
7528 assembler for grouping arithmetic expressions. If not overridden, they
7529 default to normal parentheses, which is correct for most assemblers.
7532 These macros are provided by @file{real.h} for writing the definitions
7533 of @code{ASM_OUTPUT_DOUBLE} and the like:
7535 @defmac REAL_VALUE_TO_TARGET_SINGLE (@var{x}, @var{l})
7536 @defmacx REAL_VALUE_TO_TARGET_DOUBLE (@var{x}, @var{l})
7537 @defmacx REAL_VALUE_TO_TARGET_LONG_DOUBLE (@var{x}, @var{l})
7538 @defmacx REAL_VALUE_TO_TARGET_DECIMAL32 (@var{x}, @var{l})
7539 @defmacx REAL_VALUE_TO_TARGET_DECIMAL64 (@var{x}, @var{l})
7540 @defmacx REAL_VALUE_TO_TARGET_DECIMAL128 (@var{x}, @var{l})
7541 These translate @var{x}, of type @code{REAL_VALUE_TYPE}, to the
7542 target's floating point representation, and store its bit pattern in
7543 the variable @var{l}. For @code{REAL_VALUE_TO_TARGET_SINGLE} and
7544 @code{REAL_VALUE_TO_TARGET_DECIMAL32}, this variable should be a
7545 simple @code{long int}. For the others, it should be an array of
7546 @code{long int}. The number of elements in this array is determined
7547 by the size of the desired target floating point data type: 32 bits of
7548 it go in each @code{long int} array element. Each array element holds
7549 32 bits of the result, even if @code{long int} is wider than 32 bits
7550 on the host machine.
7552 The array element values are designed so that you can print them out
7553 using @code{fprintf} in the order they should appear in the target
7557 @node Uninitialized Data
7558 @subsection Output of Uninitialized Variables
7560 Each of the macros in this section is used to do the whole job of
7561 outputting a single uninitialized variable.
7563 @defmac ASM_OUTPUT_COMMON (@var{stream}, @var{name}, @var{size}, @var{rounded})
7564 A C statement (sans semicolon) to output to the stdio stream
7565 @var{stream} the assembler definition of a common-label named
7566 @var{name} whose size is @var{size} bytes. The variable @var{rounded}
7567 is the size rounded up to whatever alignment the caller wants. It is
7568 possible that @var{size} may be zero, for instance if a struct with no
7569 other member than a zero-length array is defined. In this case, the
7570 backend must output a symbol definition that allocates at least one
7571 byte, both so that the address of the resulting object does not compare
7572 equal to any other, and because some object formats cannot even express
7573 the concept of a zero-sized common symbol, as that is how they represent
7574 an ordinary undefined external.
7576 Use the expression @code{assemble_name (@var{stream}, @var{name})} to
7577 output the name itself; before and after that, output the additional
7578 assembler syntax for defining the name, and a newline.
7580 This macro controls how the assembler definitions of uninitialized
7581 common global variables are output.
7584 @defmac ASM_OUTPUT_ALIGNED_COMMON (@var{stream}, @var{name}, @var{size}, @var{alignment})
7585 Like @code{ASM_OUTPUT_COMMON} except takes the required alignment as a
7586 separate, explicit argument. If you define this macro, it is used in
7587 place of @code{ASM_OUTPUT_COMMON}, and gives you more flexibility in
7588 handling the required alignment of the variable. The alignment is specified
7589 as the number of bits.
7592 @defmac ASM_OUTPUT_ALIGNED_DECL_COMMON (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{alignment})
7593 Like @code{ASM_OUTPUT_ALIGNED_COMMON} except that @var{decl} of the
7594 variable to be output, if there is one, or @code{NULL_TREE} if there
7595 is no corresponding variable. If you define this macro, GCC will use it
7596 in place of both @code{ASM_OUTPUT_COMMON} and
7597 @code{ASM_OUTPUT_ALIGNED_COMMON}. Define this macro when you need to see
7598 the variable's decl in order to chose what to output.
7601 @defmac ASM_OUTPUT_ALIGNED_BSS (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{alignment})
7602 A C statement (sans semicolon) to output to the stdio stream
7603 @var{stream} the assembler definition of uninitialized global @var{decl} named
7604 @var{name} whose size is @var{size} bytes. The variable @var{alignment}
7605 is the alignment specified as the number of bits.
7607 Try to use function @code{asm_output_aligned_bss} defined in file
7608 @file{varasm.c} when defining this macro. If unable, use the expression
7609 @code{assemble_name (@var{stream}, @var{name})} to output the name itself;
7610 before and after that, output the additional assembler syntax for defining
7611 the name, and a newline.
7613 There are two ways of handling global BSS@. One is to define this macro.
7614 The other is to have @code{TARGET_ASM_SELECT_SECTION} return a
7615 switchable BSS section (@pxref{TARGET_HAVE_SWITCHABLE_BSS_SECTIONS}).
7616 You do not need to do both.
7618 Some languages do not have @code{common} data, and require a
7619 non-common form of global BSS in order to handle uninitialized globals
7620 efficiently. C++ is one example of this. However, if the target does
7621 not support global BSS, the front end may choose to make globals
7622 common in order to save space in the object file.
7625 @defmac ASM_OUTPUT_LOCAL (@var{stream}, @var{name}, @var{size}, @var{rounded})
7626 A C statement (sans semicolon) to output to the stdio stream
7627 @var{stream} the assembler definition of a local-common-label named
7628 @var{name} whose size is @var{size} bytes. The variable @var{rounded}
7629 is the size rounded up to whatever alignment the caller wants.
7631 Use the expression @code{assemble_name (@var{stream}, @var{name})} to
7632 output the name itself; before and after that, output the additional
7633 assembler syntax for defining the name, and a newline.
7635 This macro controls how the assembler definitions of uninitialized
7636 static variables are output.
7639 @defmac ASM_OUTPUT_ALIGNED_LOCAL (@var{stream}, @var{name}, @var{size}, @var{alignment})
7640 Like @code{ASM_OUTPUT_LOCAL} except takes the required alignment as a
7641 separate, explicit argument. If you define this macro, it is used in
7642 place of @code{ASM_OUTPUT_LOCAL}, and gives you more flexibility in
7643 handling the required alignment of the variable. The alignment is specified
7644 as the number of bits.
7647 @defmac ASM_OUTPUT_ALIGNED_DECL_LOCAL (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{alignment})
7648 Like @code{ASM_OUTPUT_ALIGNED_DECL} except that @var{decl} of the
7649 variable to be output, if there is one, or @code{NULL_TREE} if there
7650 is no corresponding variable. If you define this macro, GCC will use it
7651 in place of both @code{ASM_OUTPUT_DECL} and
7652 @code{ASM_OUTPUT_ALIGNED_DECL}. Define this macro when you need to see
7653 the variable's decl in order to chose what to output.
7657 @subsection Output and Generation of Labels
7659 @c prevent bad page break with this line
7660 This is about outputting labels.
7662 @findex assemble_name
7663 @defmac ASM_OUTPUT_LABEL (@var{stream}, @var{name})
7664 A C statement (sans semicolon) to output to the stdio stream
7665 @var{stream} the assembler definition of a label named @var{name}.
7666 Use the expression @code{assemble_name (@var{stream}, @var{name})} to
7667 output the name itself; before and after that, output the additional
7668 assembler syntax for defining the name, and a newline. A default
7669 definition of this macro is provided which is correct for most systems.
7672 @defmac ASM_OUTPUT_FUNCTION_LABEL (@var{stream}, @var{name}, @var{decl})
7673 A C statement (sans semicolon) to output to the stdio stream
7674 @var{stream} the assembler definition of a label named @var{name} of
7676 Use the expression @code{assemble_name (@var{stream}, @var{name})} to
7677 output the name itself; before and after that, output the additional
7678 assembler syntax for defining the name, and a newline. A default
7679 definition of this macro is provided which is correct for most systems.
7681 If this macro is not defined, then the function name is defined in the
7682 usual manner as a label (by means of @code{ASM_OUTPUT_LABEL}).
7685 @findex assemble_name_raw
7686 @defmac ASM_OUTPUT_INTERNAL_LABEL (@var{stream}, @var{name})
7687 Identical to @code{ASM_OUTPUT_LABEL}, except that @var{name} is known
7688 to refer to a compiler-generated label. The default definition uses
7689 @code{assemble_name_raw}, which is like @code{assemble_name} except
7690 that it is more efficient.
7694 A C string containing the appropriate assembler directive to specify the
7695 size of a symbol, without any arguments. On systems that use ELF, the
7696 default (in @file{config/elfos.h}) is @samp{"\t.size\t"}; on other
7697 systems, the default is not to define this macro.
7699 Define this macro only if it is correct to use the default definitions
7700 of @code{ASM_OUTPUT_SIZE_DIRECTIVE} and @code{ASM_OUTPUT_MEASURED_SIZE}
7701 for your system. If you need your own custom definitions of those
7702 macros, or if you do not need explicit symbol sizes at all, do not
7706 @defmac ASM_OUTPUT_SIZE_DIRECTIVE (@var{stream}, @var{name}, @var{size})
7707 A C statement (sans semicolon) to output to the stdio stream
7708 @var{stream} a directive telling the assembler that the size of the
7709 symbol @var{name} is @var{size}. @var{size} is a @code{HOST_WIDE_INT}.
7710 If you define @code{SIZE_ASM_OP}, a default definition of this macro is
7714 @defmac ASM_OUTPUT_MEASURED_SIZE (@var{stream}, @var{name})
7715 A C statement (sans semicolon) to output to the stdio stream
7716 @var{stream} a directive telling the assembler to calculate the size of
7717 the symbol @var{name} by subtracting its address from the current
7720 If you define @code{SIZE_ASM_OP}, a default definition of this macro is
7721 provided. The default assumes that the assembler recognizes a special
7722 @samp{.} symbol as referring to the current address, and can calculate
7723 the difference between this and another symbol. If your assembler does
7724 not recognize @samp{.} or cannot do calculations with it, you will need
7725 to redefine @code{ASM_OUTPUT_MEASURED_SIZE} to use some other technique.
7729 A C string containing the appropriate assembler directive to specify the
7730 type of a symbol, without any arguments. On systems that use ELF, the
7731 default (in @file{config/elfos.h}) is @samp{"\t.type\t"}; on other
7732 systems, the default is not to define this macro.
7734 Define this macro only if it is correct to use the default definition of
7735 @code{ASM_OUTPUT_TYPE_DIRECTIVE} for your system. If you need your own
7736 custom definition of this macro, or if you do not need explicit symbol
7737 types at all, do not define this macro.
7740 @defmac TYPE_OPERAND_FMT
7741 A C string which specifies (using @code{printf} syntax) the format of
7742 the second operand to @code{TYPE_ASM_OP}. On systems that use ELF, the
7743 default (in @file{config/elfos.h}) is @samp{"@@%s"}; on other systems,
7744 the default is not to define this macro.
7746 Define this macro only if it is correct to use the default definition of
7747 @code{ASM_OUTPUT_TYPE_DIRECTIVE} for your system. If you need your own
7748 custom definition of this macro, or if you do not need explicit symbol
7749 types at all, do not define this macro.
7752 @defmac ASM_OUTPUT_TYPE_DIRECTIVE (@var{stream}, @var{type})
7753 A C statement (sans semicolon) to output to the stdio stream
7754 @var{stream} a directive telling the assembler that the type of the
7755 symbol @var{name} is @var{type}. @var{type} is a C string; currently,
7756 that string is always either @samp{"function"} or @samp{"object"}, but
7757 you should not count on this.
7759 If you define @code{TYPE_ASM_OP} and @code{TYPE_OPERAND_FMT}, a default
7760 definition of this macro is provided.
7763 @defmac ASM_DECLARE_FUNCTION_NAME (@var{stream}, @var{name}, @var{decl})
7764 A C statement (sans semicolon) to output to the stdio stream
7765 @var{stream} any text necessary for declaring the name @var{name} of a
7766 function which is being defined. This macro is responsible for
7767 outputting the label definition (perhaps using
7768 @code{ASM_OUTPUT_FUNCTION_LABEL}). The argument @var{decl} is the
7769 @code{FUNCTION_DECL} tree node representing the function.
7771 If this macro is not defined, then the function name is defined in the
7772 usual manner as a label (by means of @code{ASM_OUTPUT_FUNCTION_LABEL}).
7774 You may wish to use @code{ASM_OUTPUT_TYPE_DIRECTIVE} in the definition
7778 @defmac ASM_DECLARE_FUNCTION_SIZE (@var{stream}, @var{name}, @var{decl})
7779 A C statement (sans semicolon) to output to the stdio stream
7780 @var{stream} any text necessary for declaring the size of a function
7781 which is being defined. The argument @var{name} is the name of the
7782 function. The argument @var{decl} is the @code{FUNCTION_DECL} tree node
7783 representing the function.
7785 If this macro is not defined, then the function size is not defined.
7787 You may wish to use @code{ASM_OUTPUT_MEASURED_SIZE} in the definition
7791 @defmac ASM_DECLARE_OBJECT_NAME (@var{stream}, @var{name}, @var{decl})
7792 A C statement (sans semicolon) to output to the stdio stream
7793 @var{stream} any text necessary for declaring the name @var{name} of an
7794 initialized variable which is being defined. This macro must output the
7795 label definition (perhaps using @code{ASM_OUTPUT_LABEL}). The argument
7796 @var{decl} is the @code{VAR_DECL} tree node representing the variable.
7798 If this macro is not defined, then the variable name is defined in the
7799 usual manner as a label (by means of @code{ASM_OUTPUT_LABEL}).
7801 You may wish to use @code{ASM_OUTPUT_TYPE_DIRECTIVE} and/or
7802 @code{ASM_OUTPUT_SIZE_DIRECTIVE} in the definition of this macro.
7805 @hook TARGET_ASM_DECLARE_CONSTANT_NAME
7806 A target hook to output to the stdio stream @var{file} any text necessary
7807 for declaring the name @var{name} of a constant which is being defined. This
7808 target hook is responsible for outputting the label definition (perhaps using
7809 @code{assemble_label}). The argument @var{exp} is the value of the constant,
7810 and @var{size} is the size of the constant in bytes. The @var{name}
7811 will be an internal label.
7813 The default version of this target hook, define the @var{name} in the
7814 usual manner as a label (by means of @code{assemble_label}).
7816 You may wish to use @code{ASM_OUTPUT_TYPE_DIRECTIVE} in this target hook.
7819 @defmac ASM_DECLARE_REGISTER_GLOBAL (@var{stream}, @var{decl}, @var{regno}, @var{name})
7820 A C statement (sans semicolon) to output to the stdio stream
7821 @var{stream} any text necessary for claiming a register @var{regno}
7822 for a global variable @var{decl} with name @var{name}.
7824 If you don't define this macro, that is equivalent to defining it to do
7828 @defmac ASM_FINISH_DECLARE_OBJECT (@var{stream}, @var{decl}, @var{toplevel}, @var{atend})
7829 A C statement (sans semicolon) to finish up declaring a variable name
7830 once the compiler has processed its initializer fully and thus has had a
7831 chance to determine the size of an array when controlled by an
7832 initializer. This is used on systems where it's necessary to declare
7833 something about the size of the object.
7835 If you don't define this macro, that is equivalent to defining it to do
7838 You may wish to use @code{ASM_OUTPUT_SIZE_DIRECTIVE} and/or
7839 @code{ASM_OUTPUT_MEASURED_SIZE} in the definition of this macro.
7842 @hook TARGET_ASM_GLOBALIZE_LABEL
7843 This target hook is a function to output to the stdio stream
7844 @var{stream} some commands that will make the label @var{name} global;
7845 that is, available for reference from other files.
7847 The default implementation relies on a proper definition of
7848 @code{GLOBAL_ASM_OP}.
7851 @hook TARGET_ASM_GLOBALIZE_DECL_NAME
7852 This target hook is a function to output to the stdio stream
7853 @var{stream} some commands that will make the name associated with @var{decl}
7854 global; that is, available for reference from other files.
7856 The default implementation uses the TARGET_ASM_GLOBALIZE_LABEL target hook.
7859 @defmac ASM_WEAKEN_LABEL (@var{stream}, @var{name})
7860 A C statement (sans semicolon) to output to the stdio stream
7861 @var{stream} some commands that will make the label @var{name} weak;
7862 that is, available for reference from other files but only used if
7863 no other definition is available. Use the expression
7864 @code{assemble_name (@var{stream}, @var{name})} to output the name
7865 itself; before and after that, output the additional assembler syntax
7866 for making that name weak, and a newline.
7868 If you don't define this macro or @code{ASM_WEAKEN_DECL}, GCC will not
7869 support weak symbols and you should not define the @code{SUPPORTS_WEAK}
7873 @defmac ASM_WEAKEN_DECL (@var{stream}, @var{decl}, @var{name}, @var{value})
7874 Combines (and replaces) the function of @code{ASM_WEAKEN_LABEL} and
7875 @code{ASM_OUTPUT_WEAK_ALIAS}, allowing access to the associated function
7876 or variable decl. If @var{value} is not @code{NULL}, this C statement
7877 should output to the stdio stream @var{stream} assembler code which
7878 defines (equates) the weak symbol @var{name} to have the value
7879 @var{value}. If @var{value} is @code{NULL}, it should output commands
7880 to make @var{name} weak.
7883 @defmac ASM_OUTPUT_WEAKREF (@var{stream}, @var{decl}, @var{name}, @var{value})
7884 Outputs a directive that enables @var{name} to be used to refer to
7885 symbol @var{value} with weak-symbol semantics. @code{decl} is the
7886 declaration of @code{name}.
7889 @defmac SUPPORTS_WEAK
7890 A preprocessor constant expression which evaluates to true if the target
7891 supports weak symbols.
7893 If you don't define this macro, @file{defaults.h} provides a default
7894 definition. If either @code{ASM_WEAKEN_LABEL} or @code{ASM_WEAKEN_DECL}
7895 is defined, the default definition is @samp{1}; otherwise, it is @samp{0}.
7898 @defmac TARGET_SUPPORTS_WEAK
7899 A C expression which evaluates to true if the target supports weak symbols.
7901 If you don't define this macro, @file{defaults.h} provides a default
7902 definition. The default definition is @samp{(SUPPORTS_WEAK)}. Define
7903 this macro if you want to control weak symbol support with a compiler
7904 flag such as @option{-melf}.
7907 @defmac MAKE_DECL_ONE_ONLY (@var{decl})
7908 A C statement (sans semicolon) to mark @var{decl} to be emitted as a
7909 public symbol such that extra copies in multiple translation units will
7910 be discarded by the linker. Define this macro if your object file
7911 format provides support for this concept, such as the @samp{COMDAT}
7912 section flags in the Microsoft Windows PE/COFF format, and this support
7913 requires changes to @var{decl}, such as putting it in a separate section.
7916 @defmac SUPPORTS_ONE_ONLY
7917 A C expression which evaluates to true if the target supports one-only
7920 If you don't define this macro, @file{varasm.c} provides a default
7921 definition. If @code{MAKE_DECL_ONE_ONLY} is defined, the default
7922 definition is @samp{1}; otherwise, it is @samp{0}. Define this macro if
7923 you want to control one-only symbol support with a compiler flag, or if
7924 setting the @code{DECL_ONE_ONLY} flag is enough to mark a declaration to
7925 be emitted as one-only.
7928 @hook TARGET_ASM_ASSEMBLE_VISIBILITY
7929 This target hook is a function to output to @var{asm_out_file} some
7930 commands that will make the symbol(s) associated with @var{decl} have
7931 hidden, protected or internal visibility as specified by @var{visibility}.
7934 @defmac TARGET_WEAK_NOT_IN_ARCHIVE_TOC
7935 A C expression that evaluates to true if the target's linker expects
7936 that weak symbols do not appear in a static archive's table of contents.
7937 The default is @code{0}.
7939 Leaving weak symbols out of an archive's table of contents means that,
7940 if a symbol will only have a definition in one translation unit and
7941 will have undefined references from other translation units, that
7942 symbol should not be weak. Defining this macro to be nonzero will
7943 thus have the effect that certain symbols that would normally be weak
7944 (explicit template instantiations, and vtables for polymorphic classes
7945 with noninline key methods) will instead be nonweak.
7947 The C++ ABI requires this macro to be zero. Define this macro for
7948 targets where full C++ ABI compliance is impossible and where linker
7949 restrictions require weak symbols to be left out of a static archive's
7953 @defmac ASM_OUTPUT_EXTERNAL (@var{stream}, @var{decl}, @var{name})
7954 A C statement (sans semicolon) to output to the stdio stream
7955 @var{stream} any text necessary for declaring the name of an external
7956 symbol named @var{name} which is referenced in this compilation but
7957 not defined. The value of @var{decl} is the tree node for the
7960 This macro need not be defined if it does not need to output anything.
7961 The GNU assembler and most Unix assemblers don't require anything.
7964 @hook TARGET_ASM_EXTERNAL_LIBCALL
7965 This target hook is a function to output to @var{asm_out_file} an assembler
7966 pseudo-op to declare a library function name external. The name of the
7967 library function is given by @var{symref}, which is a @code{symbol_ref}.
7970 @hook TARGET_ASM_MARK_DECL_PRESERVED
7971 This target hook is a function to output to @var{asm_out_file} an assembler
7972 directive to annotate @var{symbol} as used. The Darwin target uses the
7973 .no_dead_code_strip directive.
7976 @defmac ASM_OUTPUT_LABELREF (@var{stream}, @var{name})
7977 A C statement (sans semicolon) to output to the stdio stream
7978 @var{stream} a reference in assembler syntax to a label named
7979 @var{name}. This should add @samp{_} to the front of the name, if that
7980 is customary on your operating system, as it is in most Berkeley Unix
7981 systems. This macro is used in @code{assemble_name}.
7984 @hook TARGET_MANGLE_ASSEMBLER_NAME
7986 @defmac ASM_OUTPUT_SYMBOL_REF (@var{stream}, @var{sym})
7987 A C statement (sans semicolon) to output a reference to
7988 @code{SYMBOL_REF} @var{sym}. If not defined, @code{assemble_name}
7989 will be used to output the name of the symbol. This macro may be used
7990 to modify the way a symbol is referenced depending on information
7991 encoded by @code{TARGET_ENCODE_SECTION_INFO}.
7994 @defmac ASM_OUTPUT_LABEL_REF (@var{stream}, @var{buf})
7995 A C statement (sans semicolon) to output a reference to @var{buf}, the
7996 result of @code{ASM_GENERATE_INTERNAL_LABEL}. If not defined,
7997 @code{assemble_name} will be used to output the name of the symbol.
7998 This macro is not used by @code{output_asm_label}, or the @code{%l}
7999 specifier that calls it; the intention is that this macro should be set
8000 when it is necessary to output a label differently when its address is
8004 @hook TARGET_ASM_INTERNAL_LABEL
8005 A function to output to the stdio stream @var{stream} a label whose
8006 name is made from the string @var{prefix} and the number @var{labelno}.
8008 It is absolutely essential that these labels be distinct from the labels
8009 used for user-level functions and variables. Otherwise, certain programs
8010 will have name conflicts with internal labels.
8012 It is desirable to exclude internal labels from the symbol table of the
8013 object file. Most assemblers have a naming convention for labels that
8014 should be excluded; on many systems, the letter @samp{L} at the
8015 beginning of a label has this effect. You should find out what
8016 convention your system uses, and follow it.
8018 The default version of this function utilizes @code{ASM_GENERATE_INTERNAL_LABEL}.
8021 @defmac ASM_OUTPUT_DEBUG_LABEL (@var{stream}, @var{prefix}, @var{num})
8022 A C statement to output to the stdio stream @var{stream} a debug info
8023 label whose name is made from the string @var{prefix} and the number
8024 @var{num}. This is useful for VLIW targets, where debug info labels
8025 may need to be treated differently than branch target labels. On some
8026 systems, branch target labels must be at the beginning of instruction
8027 bundles, but debug info labels can occur in the middle of instruction
8030 If this macro is not defined, then @code{(*targetm.asm_out.internal_label)} will be
8034 @defmac ASM_GENERATE_INTERNAL_LABEL (@var{string}, @var{prefix}, @var{num})
8035 A C statement to store into the string @var{string} a label whose name
8036 is made from the string @var{prefix} and the number @var{num}.
8038 This string, when output subsequently by @code{assemble_name}, should
8039 produce the output that @code{(*targetm.asm_out.internal_label)} would produce
8040 with the same @var{prefix} and @var{num}.
8042 If the string begins with @samp{*}, then @code{assemble_name} will
8043 output the rest of the string unchanged. It is often convenient for
8044 @code{ASM_GENERATE_INTERNAL_LABEL} to use @samp{*} in this way. If the
8045 string doesn't start with @samp{*}, then @code{ASM_OUTPUT_LABELREF} gets
8046 to output the string, and may change it. (Of course,
8047 @code{ASM_OUTPUT_LABELREF} is also part of your machine description, so
8048 you should know what it does on your machine.)
8051 @defmac ASM_FORMAT_PRIVATE_NAME (@var{outvar}, @var{name}, @var{number})
8052 A C expression to assign to @var{outvar} (which is a variable of type
8053 @code{char *}) a newly allocated string made from the string
8054 @var{name} and the number @var{number}, with some suitable punctuation
8055 added. Use @code{alloca} to get space for the string.
8057 The string will be used as an argument to @code{ASM_OUTPUT_LABELREF} to
8058 produce an assembler label for an internal static variable whose name is
8059 @var{name}. Therefore, the string must be such as to result in valid
8060 assembler code. The argument @var{number} is different each time this
8061 macro is executed; it prevents conflicts between similarly-named
8062 internal static variables in different scopes.
8064 Ideally this string should not be a valid C identifier, to prevent any
8065 conflict with the user's own symbols. Most assemblers allow periods
8066 or percent signs in assembler symbols; putting at least one of these
8067 between the name and the number will suffice.
8069 If this macro is not defined, a default definition will be provided
8070 which is correct for most systems.
8073 @defmac ASM_OUTPUT_DEF (@var{stream}, @var{name}, @var{value})
8074 A C statement to output to the stdio stream @var{stream} assembler code
8075 which defines (equates) the symbol @var{name} to have the value @var{value}.
8078 If @code{SET_ASM_OP} is defined, a default definition is provided which is
8079 correct for most systems.
8082 @defmac ASM_OUTPUT_DEF_FROM_DECLS (@var{stream}, @var{decl_of_name}, @var{decl_of_value})
8083 A C statement to output to the stdio stream @var{stream} assembler code
8084 which defines (equates) the symbol whose tree node is @var{decl_of_name}
8085 to have the value of the tree node @var{decl_of_value}. This macro will
8086 be used in preference to @samp{ASM_OUTPUT_DEF} if it is defined and if
8087 the tree nodes are available.
8090 If @code{SET_ASM_OP} is defined, a default definition is provided which is
8091 correct for most systems.
8094 @defmac TARGET_DEFERRED_OUTPUT_DEFS (@var{decl_of_name}, @var{decl_of_value})
8095 A C statement that evaluates to true if the assembler code which defines
8096 (equates) the symbol whose tree node is @var{decl_of_name} to have the value
8097 of the tree node @var{decl_of_value} should be emitted near the end of the
8098 current compilation unit. The default is to not defer output of defines.
8099 This macro affects defines output by @samp{ASM_OUTPUT_DEF} and
8100 @samp{ASM_OUTPUT_DEF_FROM_DECLS}.
8103 @defmac ASM_OUTPUT_WEAK_ALIAS (@var{stream}, @var{name}, @var{value})
8104 A C statement to output to the stdio stream @var{stream} assembler code
8105 which defines (equates) the weak symbol @var{name} to have the value
8106 @var{value}. If @var{value} is @code{NULL}, it defines @var{name} as
8107 an undefined weak symbol.
8109 Define this macro if the target only supports weak aliases; define
8110 @code{ASM_OUTPUT_DEF} instead if possible.
8113 @defmac OBJC_GEN_METHOD_LABEL (@var{buf}, @var{is_inst}, @var{class_name}, @var{cat_name}, @var{sel_name})
8114 Define this macro to override the default assembler names used for
8115 Objective-C methods.
8117 The default name is a unique method number followed by the name of the
8118 class (e.g.@: @samp{_1_Foo}). For methods in categories, the name of
8119 the category is also included in the assembler name (e.g.@:
8122 These names are safe on most systems, but make debugging difficult since
8123 the method's selector is not present in the name. Therefore, particular
8124 systems define other ways of computing names.
8126 @var{buf} is an expression of type @code{char *} which gives you a
8127 buffer in which to store the name; its length is as long as
8128 @var{class_name}, @var{cat_name} and @var{sel_name} put together, plus
8129 50 characters extra.
8131 The argument @var{is_inst} specifies whether the method is an instance
8132 method or a class method; @var{class_name} is the name of the class;
8133 @var{cat_name} is the name of the category (or @code{NULL} if the method is not
8134 in a category); and @var{sel_name} is the name of the selector.
8136 On systems where the assembler can handle quoted names, you can use this
8137 macro to provide more human-readable names.
8140 @defmac ASM_DECLARE_CLASS_REFERENCE (@var{stream}, @var{name})
8141 A C statement (sans semicolon) to output to the stdio stream
8142 @var{stream} commands to declare that the label @var{name} is an
8143 Objective-C class reference. This is only needed for targets whose
8144 linkers have special support for NeXT-style runtimes.
8147 @defmac ASM_DECLARE_UNRESOLVED_REFERENCE (@var{stream}, @var{name})
8148 A C statement (sans semicolon) to output to the stdio stream
8149 @var{stream} commands to declare that the label @var{name} is an
8150 unresolved Objective-C class reference. This is only needed for targets
8151 whose linkers have special support for NeXT-style runtimes.
8154 @node Initialization
8155 @subsection How Initialization Functions Are Handled
8156 @cindex initialization routines
8157 @cindex termination routines
8158 @cindex constructors, output of
8159 @cindex destructors, output of
8161 The compiled code for certain languages includes @dfn{constructors}
8162 (also called @dfn{initialization routines})---functions to initialize
8163 data in the program when the program is started. These functions need
8164 to be called before the program is ``started''---that is to say, before
8165 @code{main} is called.
8167 Compiling some languages generates @dfn{destructors} (also called
8168 @dfn{termination routines}) that should be called when the program
8171 To make the initialization and termination functions work, the compiler
8172 must output something in the assembler code to cause those functions to
8173 be called at the appropriate time. When you port the compiler to a new
8174 system, you need to specify how to do this.
8176 There are two major ways that GCC currently supports the execution of
8177 initialization and termination functions. Each way has two variants.
8178 Much of the structure is common to all four variations.
8180 @findex __CTOR_LIST__
8181 @findex __DTOR_LIST__
8182 The linker must build two lists of these functions---a list of
8183 initialization functions, called @code{__CTOR_LIST__}, and a list of
8184 termination functions, called @code{__DTOR_LIST__}.
8186 Each list always begins with an ignored function pointer (which may hold
8187 0, @minus{}1, or a count of the function pointers after it, depending on
8188 the environment). This is followed by a series of zero or more function
8189 pointers to constructors (or destructors), followed by a function
8190 pointer containing zero.
8192 Depending on the operating system and its executable file format, either
8193 @file{crtstuff.c} or @file{libgcc2.c} traverses these lists at startup
8194 time and exit time. Constructors are called in reverse order of the
8195 list; destructors in forward order.
8197 The best way to handle static constructors works only for object file
8198 formats which provide arbitrarily-named sections. A section is set
8199 aside for a list of constructors, and another for a list of destructors.
8200 Traditionally these are called @samp{.ctors} and @samp{.dtors}. Each
8201 object file that defines an initialization function also puts a word in
8202 the constructor section to point to that function. The linker
8203 accumulates all these words into one contiguous @samp{.ctors} section.
8204 Termination functions are handled similarly.
8206 This method will be chosen as the default by @file{target-def.h} if
8207 @code{TARGET_ASM_NAMED_SECTION} is defined. A target that does not
8208 support arbitrary sections, but does support special designated
8209 constructor and destructor sections may define @code{CTORS_SECTION_ASM_OP}
8210 and @code{DTORS_SECTION_ASM_OP} to achieve the same effect.
8212 When arbitrary sections are available, there are two variants, depending
8213 upon how the code in @file{crtstuff.c} is called. On systems that
8214 support a @dfn{.init} section which is executed at program startup,
8215 parts of @file{crtstuff.c} are compiled into that section. The
8216 program is linked by the @command{gcc} driver like this:
8219 ld -o @var{output_file} crti.o crtbegin.o @dots{} -lgcc crtend.o crtn.o
8222 The prologue of a function (@code{__init}) appears in the @code{.init}
8223 section of @file{crti.o}; the epilogue appears in @file{crtn.o}. Likewise
8224 for the function @code{__fini} in the @dfn{.fini} section. Normally these
8225 files are provided by the operating system or by the GNU C library, but
8226 are provided by GCC for a few targets.
8228 The objects @file{crtbegin.o} and @file{crtend.o} are (for most targets)
8229 compiled from @file{crtstuff.c}. They contain, among other things, code
8230 fragments within the @code{.init} and @code{.fini} sections that branch
8231 to routines in the @code{.text} section. The linker will pull all parts
8232 of a section together, which results in a complete @code{__init} function
8233 that invokes the routines we need at startup.
8235 To use this variant, you must define the @code{INIT_SECTION_ASM_OP}
8238 If no init section is available, when GCC compiles any function called
8239 @code{main} (or more accurately, any function designated as a program
8240 entry point by the language front end calling @code{expand_main_function}),
8241 it inserts a procedure call to @code{__main} as the first executable code
8242 after the function prologue. The @code{__main} function is defined
8243 in @file{libgcc2.c} and runs the global constructors.
8245 In file formats that don't support arbitrary sections, there are again
8246 two variants. In the simplest variant, the GNU linker (GNU @code{ld})
8247 and an `a.out' format must be used. In this case,
8248 @code{TARGET_ASM_CONSTRUCTOR} is defined to produce a @code{.stabs}
8249 entry of type @samp{N_SETT}, referencing the name @code{__CTOR_LIST__},
8250 and with the address of the void function containing the initialization
8251 code as its value. The GNU linker recognizes this as a request to add
8252 the value to a @dfn{set}; the values are accumulated, and are eventually
8253 placed in the executable as a vector in the format described above, with
8254 a leading (ignored) count and a trailing zero element.
8255 @code{TARGET_ASM_DESTRUCTOR} is handled similarly. Since no init
8256 section is available, the absence of @code{INIT_SECTION_ASM_OP} causes
8257 the compilation of @code{main} to call @code{__main} as above, starting
8258 the initialization process.
8260 The last variant uses neither arbitrary sections nor the GNU linker.
8261 This is preferable when you want to do dynamic linking and when using
8262 file formats which the GNU linker does not support, such as `ECOFF'@. In
8263 this case, @code{TARGET_HAVE_CTORS_DTORS} is false, initialization and
8264 termination functions are recognized simply by their names. This requires
8265 an extra program in the linkage step, called @command{collect2}. This program
8266 pretends to be the linker, for use with GCC; it does its job by running
8267 the ordinary linker, but also arranges to include the vectors of
8268 initialization and termination functions. These functions are called
8269 via @code{__main} as described above. In order to use this method,
8270 @code{use_collect2} must be defined in the target in @file{config.gcc}.
8273 The following section describes the specific macros that control and
8274 customize the handling of initialization and termination functions.
8277 @node Macros for Initialization
8278 @subsection Macros Controlling Initialization Routines
8280 Here are the macros that control how the compiler handles initialization
8281 and termination functions:
8283 @defmac INIT_SECTION_ASM_OP
8284 If defined, a C string constant, including spacing, for the assembler
8285 operation to identify the following data as initialization code. If not
8286 defined, GCC will assume such a section does not exist. When you are
8287 using special sections for initialization and termination functions, this
8288 macro also controls how @file{crtstuff.c} and @file{libgcc2.c} arrange to
8289 run the initialization functions.
8292 @defmac HAS_INIT_SECTION
8293 If defined, @code{main} will not call @code{__main} as described above.
8294 This macro should be defined for systems that control start-up code
8295 on a symbol-by-symbol basis, such as OSF/1, and should not
8296 be defined explicitly for systems that support @code{INIT_SECTION_ASM_OP}.
8299 @defmac LD_INIT_SWITCH
8300 If defined, a C string constant for a switch that tells the linker that
8301 the following symbol is an initialization routine.
8304 @defmac LD_FINI_SWITCH
8305 If defined, a C string constant for a switch that tells the linker that
8306 the following symbol is a finalization routine.
8309 @defmac COLLECT_SHARED_INIT_FUNC (@var{stream}, @var{func})
8310 If defined, a C statement that will write a function that can be
8311 automatically called when a shared library is loaded. The function
8312 should call @var{func}, which takes no arguments. If not defined, and
8313 the object format requires an explicit initialization function, then a
8314 function called @code{_GLOBAL__DI} will be generated.
8316 This function and the following one are used by collect2 when linking a
8317 shared library that needs constructors or destructors, or has DWARF2
8318 exception tables embedded in the code.
8321 @defmac COLLECT_SHARED_FINI_FUNC (@var{stream}, @var{func})
8322 If defined, a C statement that will write a function that can be
8323 automatically called when a shared library is unloaded. The function
8324 should call @var{func}, which takes no arguments. If not defined, and
8325 the object format requires an explicit finalization function, then a
8326 function called @code{_GLOBAL__DD} will be generated.
8329 @defmac INVOKE__main
8330 If defined, @code{main} will call @code{__main} despite the presence of
8331 @code{INIT_SECTION_ASM_OP}. This macro should be defined for systems
8332 where the init section is not actually run automatically, but is still
8333 useful for collecting the lists of constructors and destructors.
8336 @defmac SUPPORTS_INIT_PRIORITY
8337 If nonzero, the C++ @code{init_priority} attribute is supported and the
8338 compiler should emit instructions to control the order of initialization
8339 of objects. If zero, the compiler will issue an error message upon
8340 encountering an @code{init_priority} attribute.
8343 @hook TARGET_HAVE_CTORS_DTORS
8344 This value is true if the target supports some ``native'' method of
8345 collecting constructors and destructors to be run at startup and exit.
8346 It is false if we must use @command{collect2}.
8349 @hook TARGET_ASM_CONSTRUCTOR
8350 If defined, a function that outputs assembler code to arrange to call
8351 the function referenced by @var{symbol} at initialization time.
8353 Assume that @var{symbol} is a @code{SYMBOL_REF} for a function taking
8354 no arguments and with no return value. If the target supports initialization
8355 priorities, @var{priority} is a value between 0 and @code{MAX_INIT_PRIORITY};
8356 otherwise it must be @code{DEFAULT_INIT_PRIORITY}.
8358 If this macro is not defined by the target, a suitable default will
8359 be chosen if (1) the target supports arbitrary section names, (2) the
8360 target defines @code{CTORS_SECTION_ASM_OP}, or (3) @code{USE_COLLECT2}
8364 @hook TARGET_ASM_DESTRUCTOR
8365 This is like @code{TARGET_ASM_CONSTRUCTOR} but used for termination
8366 functions rather than initialization functions.
8369 If @code{TARGET_HAVE_CTORS_DTORS} is true, the initialization routine
8370 generated for the generated object file will have static linkage.
8372 If your system uses @command{collect2} as the means of processing
8373 constructors, then that program normally uses @command{nm} to scan
8374 an object file for constructor functions to be called.
8376 On certain kinds of systems, you can define this macro to make
8377 @command{collect2} work faster (and, in some cases, make it work at all):
8379 @defmac OBJECT_FORMAT_COFF
8380 Define this macro if the system uses COFF (Common Object File Format)
8381 object files, so that @command{collect2} can assume this format and scan
8382 object files directly for dynamic constructor/destructor functions.
8384 This macro is effective only in a native compiler; @command{collect2} as
8385 part of a cross compiler always uses @command{nm} for the target machine.
8388 @defmac REAL_NM_FILE_NAME
8389 Define this macro as a C string constant containing the file name to use
8390 to execute @command{nm}. The default is to search the path normally for
8395 @command{collect2} calls @command{nm} to scan object files for static
8396 constructors and destructors and LTO info. By default, @option{-n} is
8397 passed. Define @code{NM_FLAGS} to a C string constant if other options
8398 are needed to get the same output format as GNU @command{nm -n}
8402 If your system supports shared libraries and has a program to list the
8403 dynamic dependencies of a given library or executable, you can define
8404 these macros to enable support for running initialization and
8405 termination functions in shared libraries:
8408 Define this macro to a C string constant containing the name of the program
8409 which lists dynamic dependencies, like @command{ldd} under SunOS 4.
8412 @defmac PARSE_LDD_OUTPUT (@var{ptr})
8413 Define this macro to be C code that extracts filenames from the output
8414 of the program denoted by @code{LDD_SUFFIX}. @var{ptr} is a variable
8415 of type @code{char *} that points to the beginning of a line of output
8416 from @code{LDD_SUFFIX}. If the line lists a dynamic dependency, the
8417 code must advance @var{ptr} to the beginning of the filename on that
8418 line. Otherwise, it must set @var{ptr} to @code{NULL}.
8421 @defmac SHLIB_SUFFIX
8422 Define this macro to a C string constant containing the default shared
8423 library extension of the target (e.g., @samp{".so"}). @command{collect2}
8424 strips version information after this suffix when generating global
8425 constructor and destructor names. This define is only needed on targets
8426 that use @command{collect2} to process constructors and destructors.
8429 @node Instruction Output
8430 @subsection Output of Assembler Instructions
8432 @c prevent bad page break with this line
8433 This describes assembler instruction output.
8435 @defmac REGISTER_NAMES
8436 A C initializer containing the assembler's names for the machine
8437 registers, each one as a C string constant. This is what translates
8438 register numbers in the compiler into assembler language.
8441 @defmac ADDITIONAL_REGISTER_NAMES
8442 If defined, a C initializer for an array of structures containing a name
8443 and a register number. This macro defines additional names for hard
8444 registers, thus allowing the @code{asm} option in declarations to refer
8445 to registers using alternate names.
8448 @defmac OVERLAPPING_REGISTER_NAMES
8449 If defined, a C initializer for an array of structures containing a
8450 name, a register number and a count of the number of consecutive
8451 machine registers the name overlaps. This macro defines additional
8452 names for hard registers, thus allowing the @code{asm} option in
8453 declarations to refer to registers using alternate names. Unlike
8454 @code{ADDITIONAL_REGISTER_NAMES}, this macro should be used when the
8455 register name implies multiple underlying registers.
8457 This macro should be used when it is important that a clobber in an
8458 @code{asm} statement clobbers all the underlying values implied by the
8459 register name. For example, on ARM, clobbering the double-precision
8460 VFP register ``d0'' implies clobbering both single-precision registers
8464 @defmac ASM_OUTPUT_OPCODE (@var{stream}, @var{ptr})
8465 Define this macro if you are using an unusual assembler that
8466 requires different names for the machine instructions.
8468 The definition is a C statement or statements which output an
8469 assembler instruction opcode to the stdio stream @var{stream}. The
8470 macro-operand @var{ptr} is a variable of type @code{char *} which
8471 points to the opcode name in its ``internal'' form---the form that is
8472 written in the machine description. The definition should output the
8473 opcode name to @var{stream}, performing any translation you desire, and
8474 increment the variable @var{ptr} to point at the end of the opcode
8475 so that it will not be output twice.
8477 In fact, your macro definition may process less than the entire opcode
8478 name, or more than the opcode name; but if you want to process text
8479 that includes @samp{%}-sequences to substitute operands, you must take
8480 care of the substitution yourself. Just be sure to increment
8481 @var{ptr} over whatever text should not be output normally.
8483 @findex recog_data.operand
8484 If you need to look at the operand values, they can be found as the
8485 elements of @code{recog_data.operand}.
8487 If the macro definition does nothing, the instruction is output
8491 @defmac FINAL_PRESCAN_INSN (@var{insn}, @var{opvec}, @var{noperands})
8492 If defined, a C statement to be executed just prior to the output of
8493 assembler code for @var{insn}, to modify the extracted operands so
8494 they will be output differently.
8496 Here the argument @var{opvec} is the vector containing the operands
8497 extracted from @var{insn}, and @var{noperands} is the number of
8498 elements of the vector which contain meaningful data for this insn.
8499 The contents of this vector are what will be used to convert the insn
8500 template into assembler code, so you can change the assembler output
8501 by changing the contents of the vector.
8503 This macro is useful when various assembler syntaxes share a single
8504 file of instruction patterns; by defining this macro differently, you
8505 can cause a large class of instructions to be output differently (such
8506 as with rearranged operands). Naturally, variations in assembler
8507 syntax affecting individual insn patterns ought to be handled by
8508 writing conditional output routines in those patterns.
8510 If this macro is not defined, it is equivalent to a null statement.
8513 @hook TARGET_ASM_FINAL_POSTSCAN_INSN
8514 If defined, this target hook is a function which is executed just after the
8515 output of assembler code for @var{insn}, to change the mode of the assembler
8518 Here the argument @var{opvec} is the vector containing the operands
8519 extracted from @var{insn}, and @var{noperands} is the number of
8520 elements of the vector which contain meaningful data for this insn.
8521 The contents of this vector are what was used to convert the insn
8522 template into assembler code, so you can change the assembler mode
8523 by checking the contents of the vector.
8526 @defmac PRINT_OPERAND (@var{stream}, @var{x}, @var{code})
8527 A C compound statement to output to stdio stream @var{stream} the
8528 assembler syntax for an instruction operand @var{x}. @var{x} is an
8531 @var{code} is a value that can be used to specify one of several ways
8532 of printing the operand. It is used when identical operands must be
8533 printed differently depending on the context. @var{code} comes from
8534 the @samp{%} specification that was used to request printing of the
8535 operand. If the specification was just @samp{%@var{digit}} then
8536 @var{code} is 0; if the specification was @samp{%@var{ltr}
8537 @var{digit}} then @var{code} is the ASCII code for @var{ltr}.
8540 If @var{x} is a register, this macro should print the register's name.
8541 The names can be found in an array @code{reg_names} whose type is
8542 @code{char *[]}. @code{reg_names} is initialized from
8543 @code{REGISTER_NAMES}.
8545 When the machine description has a specification @samp{%@var{punct}}
8546 (a @samp{%} followed by a punctuation character), this macro is called
8547 with a null pointer for @var{x} and the punctuation character for
8551 @defmac PRINT_OPERAND_PUNCT_VALID_P (@var{code})
8552 A C expression which evaluates to true if @var{code} is a valid
8553 punctuation character for use in the @code{PRINT_OPERAND} macro. If
8554 @code{PRINT_OPERAND_PUNCT_VALID_P} is not defined, it means that no
8555 punctuation characters (except for the standard one, @samp{%}) are used
8559 @defmac PRINT_OPERAND_ADDRESS (@var{stream}, @var{x})
8560 A C compound statement to output to stdio stream @var{stream} the
8561 assembler syntax for an instruction operand that is a memory reference
8562 whose address is @var{x}. @var{x} is an RTL expression.
8564 @cindex @code{TARGET_ENCODE_SECTION_INFO} usage
8565 On some machines, the syntax for a symbolic address depends on the
8566 section that the address refers to. On these machines, define the hook
8567 @code{TARGET_ENCODE_SECTION_INFO} to store the information into the
8568 @code{symbol_ref}, and then check for it here. @xref{Assembler
8572 @findex dbr_sequence_length
8573 @defmac DBR_OUTPUT_SEQEND (@var{file})
8574 A C statement, to be executed after all slot-filler instructions have
8575 been output. If necessary, call @code{dbr_sequence_length} to
8576 determine the number of slots filled in a sequence (zero if not
8577 currently outputting a sequence), to decide how many no-ops to output,
8580 Don't define this macro if it has nothing to do, but it is helpful in
8581 reading assembly output if the extent of the delay sequence is made
8582 explicit (e.g.@: with white space).
8585 @findex final_sequence
8586 Note that output routines for instructions with delay slots must be
8587 prepared to deal with not being output as part of a sequence
8588 (i.e.@: when the scheduling pass is not run, or when no slot fillers could be
8589 found.) The variable @code{final_sequence} is null when not
8590 processing a sequence, otherwise it contains the @code{sequence} rtx
8594 @defmac REGISTER_PREFIX
8595 @defmacx LOCAL_LABEL_PREFIX
8596 @defmacx USER_LABEL_PREFIX
8597 @defmacx IMMEDIATE_PREFIX
8598 If defined, C string expressions to be used for the @samp{%R}, @samp{%L},
8599 @samp{%U}, and @samp{%I} options of @code{asm_fprintf} (see
8600 @file{final.c}). These are useful when a single @file{md} file must
8601 support multiple assembler formats. In that case, the various @file{tm.h}
8602 files can define these macros differently.
8605 @defmac ASM_FPRINTF_EXTENSIONS (@var{file}, @var{argptr}, @var{format})
8606 If defined this macro should expand to a series of @code{case}
8607 statements which will be parsed inside the @code{switch} statement of
8608 the @code{asm_fprintf} function. This allows targets to define extra
8609 printf formats which may useful when generating their assembler
8610 statements. Note that uppercase letters are reserved for future
8611 generic extensions to asm_fprintf, and so are not available to target
8612 specific code. The output file is given by the parameter @var{file}.
8613 The varargs input pointer is @var{argptr} and the rest of the format
8614 string, starting the character after the one that is being switched
8615 upon, is pointed to by @var{format}.
8618 @defmac ASSEMBLER_DIALECT
8619 If your target supports multiple dialects of assembler language (such as
8620 different opcodes), define this macro as a C expression that gives the
8621 numeric index of the assembler language dialect to use, with zero as the
8624 If this macro is defined, you may use constructs of the form
8626 @samp{@{option0|option1|option2@dots{}@}}
8629 in the output templates of patterns (@pxref{Output Template}) or in the
8630 first argument of @code{asm_fprintf}. This construct outputs
8631 @samp{option0}, @samp{option1}, @samp{option2}, etc., if the value of
8632 @code{ASSEMBLER_DIALECT} is zero, one, two, etc. Any special characters
8633 within these strings retain their usual meaning. If there are fewer
8634 alternatives within the braces than the value of
8635 @code{ASSEMBLER_DIALECT}, the construct outputs nothing.
8637 If you do not define this macro, the characters @samp{@{}, @samp{|} and
8638 @samp{@}} do not have any special meaning when used in templates or
8639 operands to @code{asm_fprintf}.
8641 Define the macros @code{REGISTER_PREFIX}, @code{LOCAL_LABEL_PREFIX},
8642 @code{USER_LABEL_PREFIX} and @code{IMMEDIATE_PREFIX} if you can express
8643 the variations in assembler language syntax with that mechanism. Define
8644 @code{ASSEMBLER_DIALECT} and use the @samp{@{option0|option1@}} syntax
8645 if the syntax variant are larger and involve such things as different
8646 opcodes or operand order.
8649 @defmac ASM_OUTPUT_REG_PUSH (@var{stream}, @var{regno})
8650 A C expression to output to @var{stream} some assembler code
8651 which will push hard register number @var{regno} onto the stack.
8652 The code need not be optimal, since this macro is used only when
8656 @defmac ASM_OUTPUT_REG_POP (@var{stream}, @var{regno})
8657 A C expression to output to @var{stream} some assembler code
8658 which will pop hard register number @var{regno} off of the stack.
8659 The code need not be optimal, since this macro is used only when
8663 @node Dispatch Tables
8664 @subsection Output of Dispatch Tables
8666 @c prevent bad page break with this line
8667 This concerns dispatch tables.
8669 @cindex dispatch table
8670 @defmac ASM_OUTPUT_ADDR_DIFF_ELT (@var{stream}, @var{body}, @var{value}, @var{rel})
8671 A C statement to output to the stdio stream @var{stream} an assembler
8672 pseudo-instruction to generate a difference between two labels.
8673 @var{value} and @var{rel} are the numbers of two internal labels. The
8674 definitions of these labels are output using
8675 @code{(*targetm.asm_out.internal_label)}, and they must be printed in the same
8676 way here. For example,
8679 fprintf (@var{stream}, "\t.word L%d-L%d\n",
8680 @var{value}, @var{rel})
8683 You must provide this macro on machines where the addresses in a
8684 dispatch table are relative to the table's own address. If defined, GCC
8685 will also use this macro on all machines when producing PIC@.
8686 @var{body} is the body of the @code{ADDR_DIFF_VEC}; it is provided so that the
8687 mode and flags can be read.
8690 @defmac ASM_OUTPUT_ADDR_VEC_ELT (@var{stream}, @var{value})
8691 This macro should be provided on machines where the addresses
8692 in a dispatch table are absolute.
8694 The definition should be a C statement to output to the stdio stream
8695 @var{stream} an assembler pseudo-instruction to generate a reference to
8696 a label. @var{value} is the number of an internal label whose
8697 definition is output using @code{(*targetm.asm_out.internal_label)}.
8701 fprintf (@var{stream}, "\t.word L%d\n", @var{value})
8705 @defmac ASM_OUTPUT_CASE_LABEL (@var{stream}, @var{prefix}, @var{num}, @var{table})
8706 Define this if the label before a jump-table needs to be output
8707 specially. The first three arguments are the same as for
8708 @code{(*targetm.asm_out.internal_label)}; the fourth argument is the
8709 jump-table which follows (a @code{jump_insn} containing an
8710 @code{addr_vec} or @code{addr_diff_vec}).
8712 This feature is used on system V to output a @code{swbeg} statement
8715 If this macro is not defined, these labels are output with
8716 @code{(*targetm.asm_out.internal_label)}.
8719 @defmac ASM_OUTPUT_CASE_END (@var{stream}, @var{num}, @var{table})
8720 Define this if something special must be output at the end of a
8721 jump-table. The definition should be a C statement to be executed
8722 after the assembler code for the table is written. It should write
8723 the appropriate code to stdio stream @var{stream}. The argument
8724 @var{table} is the jump-table insn, and @var{num} is the label-number
8725 of the preceding label.
8727 If this macro is not defined, nothing special is output at the end of
8731 @hook TARGET_ASM_EMIT_UNWIND_LABEL
8732 This target hook emits a label at the beginning of each FDE@. It
8733 should be defined on targets where FDEs need special labels, and it
8734 should write the appropriate label, for the FDE associated with the
8735 function declaration @var{decl}, to the stdio stream @var{stream}.
8736 The third argument, @var{for_eh}, is a boolean: true if this is for an
8737 exception table. The fourth argument, @var{empty}, is a boolean:
8738 true if this is a placeholder label for an omitted FDE@.
8740 The default is that FDEs are not given nonlocal labels.
8743 @hook TARGET_ASM_EMIT_EXCEPT_TABLE_LABEL
8744 This target hook emits a label at the beginning of the exception table.
8745 It should be defined on targets where it is desirable for the table
8746 to be broken up according to function.
8748 The default is that no label is emitted.
8751 @hook TARGET_ASM_EMIT_EXCEPT_PERSONALITY
8753 @hook TARGET_ASM_UNWIND_EMIT
8754 This target hook emits assembly directives required to unwind the
8755 given instruction. This is only used when @code{TARGET_EXCEPT_UNWIND_INFO}
8756 returns @code{UI_TARGET}.
8759 @hook TARGET_ASM_UNWIND_EMIT_BEFORE_INSN
8761 @node Exception Region Output
8762 @subsection Assembler Commands for Exception Regions
8764 @c prevent bad page break with this line
8766 This describes commands marking the start and the end of an exception
8769 @defmac EH_FRAME_SECTION_NAME
8770 If defined, a C string constant for the name of the section containing
8771 exception handling frame unwind information. If not defined, GCC will
8772 provide a default definition if the target supports named sections.
8773 @file{crtstuff.c} uses this macro to switch to the appropriate section.
8775 You should define this symbol if your target supports DWARF 2 frame
8776 unwind information and the default definition does not work.
8779 @defmac EH_FRAME_IN_DATA_SECTION
8780 If defined, DWARF 2 frame unwind information will be placed in the
8781 data section even though the target supports named sections. This
8782 might be necessary, for instance, if the system linker does garbage
8783 collection and sections cannot be marked as not to be collected.
8785 Do not define this macro unless @code{TARGET_ASM_NAMED_SECTION} is
8789 @defmac EH_TABLES_CAN_BE_READ_ONLY
8790 Define this macro to 1 if your target is such that no frame unwind
8791 information encoding used with non-PIC code will ever require a
8792 runtime relocation, but the linker may not support merging read-only
8793 and read-write sections into a single read-write section.
8796 @defmac MASK_RETURN_ADDR
8797 An rtx used to mask the return address found via @code{RETURN_ADDR_RTX}, so
8798 that it does not contain any extraneous set bits in it.
8801 @defmac DWARF2_UNWIND_INFO
8802 Define this macro to 0 if your target supports DWARF 2 frame unwind
8803 information, but it does not yet work with exception handling.
8804 Otherwise, if your target supports this information (if it defines
8805 @code{INCOMING_RETURN_ADDR_RTX} and either @code{UNALIGNED_INT_ASM_OP}
8806 or @code{OBJECT_FORMAT_ELF}), GCC will provide a default definition of 1.
8809 @hook TARGET_EXCEPT_UNWIND_INFO
8810 This hook defines the mechanism that will be used for exception handling
8811 by the target. If the target has ABI specified unwind tables, the hook
8812 should return @code{UI_TARGET}. If the target is to use the
8813 @code{setjmp}/@code{longjmp}-based exception handling scheme, the hook
8814 should return @code{UI_SJLJ}. If the target supports DWARF 2 frame unwind
8815 information, the hook should return @code{UI_DWARF2}.
8817 A target may, if exceptions are disabled, choose to return @code{UI_NONE}.
8818 This may end up simplifying other parts of target-specific code. The
8819 default implementation of this hook never returns @code{UI_NONE}.
8821 Note that the value returned by this hook should be constant. It should
8822 not depend on anything except the command-line switches described by
8823 @var{opts}. In particular, the
8824 setting @code{UI_SJLJ} must be fixed at compiler start-up as C pre-processor
8825 macros and builtin functions related to exception handling are set up
8826 depending on this setting.
8828 The default implementation of the hook first honors the
8829 @option{--enable-sjlj-exceptions} configure option, then
8830 @code{DWARF2_UNWIND_INFO}, and finally defaults to @code{UI_SJLJ}. If
8831 @code{DWARF2_UNWIND_INFO} depends on command-line options, the target
8832 must define this hook so that @var{opts} is used correctly.
8835 @hook TARGET_UNWIND_TABLES_DEFAULT
8836 This variable should be set to @code{true} if the target ABI requires unwinding
8837 tables even when exceptions are not used. It must not be modified by
8838 command-line option processing.
8841 @defmac DONT_USE_BUILTIN_SETJMP
8842 Define this macro to 1 if the @code{setjmp}/@code{longjmp}-based scheme
8843 should use the @code{setjmp}/@code{longjmp} functions from the C library
8844 instead of the @code{__builtin_setjmp}/@code{__builtin_longjmp} machinery.
8847 @defmac DWARF_CIE_DATA_ALIGNMENT
8848 This macro need only be defined if the target might save registers in the
8849 function prologue at an offset to the stack pointer that is not aligned to
8850 @code{UNITS_PER_WORD}. The definition should be the negative minimum
8851 alignment if @code{STACK_GROWS_DOWNWARD} is defined, and the positive
8852 minimum alignment otherwise. @xref{SDB and DWARF}. Only applicable if
8853 the target supports DWARF 2 frame unwind information.
8856 @hook TARGET_TERMINATE_DW2_EH_FRAME_INFO
8857 Contains the value true if the target should add a zero word onto the
8858 end of a Dwarf-2 frame info section when used for exception handling.
8859 Default value is false if @code{EH_FRAME_SECTION_NAME} is defined, and
8863 @hook TARGET_DWARF_REGISTER_SPAN
8864 Given a register, this hook should return a parallel of registers to
8865 represent where to find the register pieces. Define this hook if the
8866 register and its mode are represented in Dwarf in non-contiguous
8867 locations, or if the register should be represented in more than one
8868 register in Dwarf. Otherwise, this hook should return @code{NULL_RTX}.
8869 If not defined, the default is to return @code{NULL_RTX}.
8872 @hook TARGET_INIT_DWARF_REG_SIZES_EXTRA
8873 If some registers are represented in Dwarf-2 unwind information in
8874 multiple pieces, define this hook to fill in information about the
8875 sizes of those pieces in the table used by the unwinder at runtime.
8876 It will be called by @code{expand_builtin_init_dwarf_reg_sizes} after
8877 filling in a single size corresponding to each hard register;
8878 @var{address} is the address of the table.
8881 @hook TARGET_ASM_TTYPE
8882 This hook is used to output a reference from a frame unwinding table to
8883 the type_info object identified by @var{sym}. It should return @code{true}
8884 if the reference was output. Returning @code{false} will cause the
8885 reference to be output using the normal Dwarf2 routines.
8888 @hook TARGET_ARM_EABI_UNWINDER
8889 This flag should be set to @code{true} on targets that use an ARM EABI
8890 based unwinding library, and @code{false} on other targets. This effects
8891 the format of unwinding tables, and how the unwinder in entered after
8892 running a cleanup. The default is @code{false}.
8895 @node Alignment Output
8896 @subsection Assembler Commands for Alignment
8898 @c prevent bad page break with this line
8899 This describes commands for alignment.
8901 @defmac JUMP_ALIGN (@var{label})
8902 The alignment (log base 2) to put in front of @var{label}, which is
8903 a common destination of jumps and has no fallthru incoming edge.
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
8909 Unless it's necessary to inspect the @var{label} parameter, it is better
8910 to set the variable @var{align_jumps} in the target's
8911 @code{TARGET_OPTION_OVERRIDE}. Otherwise, you should try to honor the user's
8912 selection in @var{align_jumps} in a @code{JUMP_ALIGN} implementation.
8915 @hook TARGET_ASM_JUMP_ALIGN_MAX_SKIP
8916 The maximum number of bytes to skip before @var{label} when applying
8917 @code{JUMP_ALIGN}. This works only if
8918 @code{ASM_OUTPUT_MAX_SKIP_ALIGN} is defined.
8921 @defmac LABEL_ALIGN_AFTER_BARRIER (@var{label})
8922 The alignment (log base 2) to put in front of @var{label}, which follows
8925 This macro need not be defined if you don't want any special alignment
8926 to be done at such a time. Most machine descriptions do not currently
8930 @hook TARGET_ASM_LABEL_ALIGN_AFTER_BARRIER_MAX_SKIP
8931 The maximum number of bytes to skip before @var{label} when applying
8932 @code{LABEL_ALIGN_AFTER_BARRIER}. This works only if
8933 @code{ASM_OUTPUT_MAX_SKIP_ALIGN} is defined.
8936 @defmac LOOP_ALIGN (@var{label})
8937 The alignment (log base 2) to put in front of @var{label}, which follows
8938 a @code{NOTE_INSN_LOOP_BEG} note.
8940 This macro need not be defined if you don't want any special alignment
8941 to be done at such a time. Most machine descriptions do not currently
8944 Unless it's necessary to inspect the @var{label} parameter, it is better
8945 to set the variable @code{align_loops} in the target's
8946 @code{TARGET_OPTION_OVERRIDE}. Otherwise, you should try to honor the user's
8947 selection in @code{align_loops} in a @code{LOOP_ALIGN} implementation.
8950 @hook TARGET_ASM_LOOP_ALIGN_MAX_SKIP
8951 The maximum number of bytes to skip when applying @code{LOOP_ALIGN} to
8952 @var{label}. This works only if @code{ASM_OUTPUT_MAX_SKIP_ALIGN} is
8956 @defmac LABEL_ALIGN (@var{label})
8957 The alignment (log base 2) to put in front of @var{label}.
8958 If @code{LABEL_ALIGN_AFTER_BARRIER} / @code{LOOP_ALIGN} specify a different alignment,
8959 the maximum of the specified values is used.
8961 Unless it's necessary to inspect the @var{label} parameter, it is better
8962 to set the variable @code{align_labels} in the target's
8963 @code{TARGET_OPTION_OVERRIDE}. Otherwise, you should try to honor the user's
8964 selection in @code{align_labels} in a @code{LABEL_ALIGN} implementation.
8967 @hook TARGET_ASM_LABEL_ALIGN_MAX_SKIP
8968 The maximum number of bytes to skip when applying @code{LABEL_ALIGN}
8969 to @var{label}. This works only if @code{ASM_OUTPUT_MAX_SKIP_ALIGN}
8973 @defmac ASM_OUTPUT_SKIP (@var{stream}, @var{nbytes})
8974 A C statement to output to the stdio stream @var{stream} an assembler
8975 instruction to advance the location counter by @var{nbytes} bytes.
8976 Those bytes should be zero when loaded. @var{nbytes} will be a C
8977 expression of type @code{unsigned HOST_WIDE_INT}.
8980 @defmac ASM_NO_SKIP_IN_TEXT
8981 Define this macro if @code{ASM_OUTPUT_SKIP} should not be used in the
8982 text section because it fails to put zeros in the bytes that are skipped.
8983 This is true on many Unix systems, where the pseudo--op to skip bytes
8984 produces no-op instructions rather than zeros when used in the text
8988 @defmac ASM_OUTPUT_ALIGN (@var{stream}, @var{power})
8989 A C statement to output to the stdio stream @var{stream} an assembler
8990 command to advance the location counter to a multiple of 2 to the
8991 @var{power} bytes. @var{power} will be a C expression of type @code{int}.
8994 @defmac ASM_OUTPUT_ALIGN_WITH_NOP (@var{stream}, @var{power})
8995 Like @code{ASM_OUTPUT_ALIGN}, except that the ``nop'' instruction is used
8996 for padding, if necessary.
8999 @defmac ASM_OUTPUT_MAX_SKIP_ALIGN (@var{stream}, @var{power}, @var{max_skip})
9000 A C statement to output to the stdio stream @var{stream} an assembler
9001 command to advance the location counter to a multiple of 2 to the
9002 @var{power} bytes, but only if @var{max_skip} or fewer bytes are needed to
9003 satisfy the alignment request. @var{power} and @var{max_skip} will be
9004 a C expression of type @code{int}.
9008 @node Debugging Info
9009 @section Controlling Debugging Information Format
9011 @c prevent bad page break with this line
9012 This describes how to specify debugging information.
9015 * All Debuggers:: Macros that affect all debugging formats uniformly.
9016 * DBX Options:: Macros enabling specific options in DBX format.
9017 * DBX Hooks:: Hook macros for varying DBX format.
9018 * File Names and DBX:: Macros controlling output of file names in DBX format.
9019 * SDB and DWARF:: Macros for SDB (COFF) and DWARF formats.
9020 * VMS Debug:: Macros for VMS debug format.
9024 @subsection Macros Affecting All Debugging Formats
9026 @c prevent bad page break with this line
9027 These macros affect all debugging formats.
9029 @defmac DBX_REGISTER_NUMBER (@var{regno})
9030 A C expression that returns the DBX register number for the compiler
9031 register number @var{regno}. In the default macro provided, the value
9032 of this expression will be @var{regno} itself. But sometimes there are
9033 some registers that the compiler knows about and DBX does not, or vice
9034 versa. In such cases, some register may need to have one number in the
9035 compiler and another for DBX@.
9037 If two registers have consecutive numbers inside GCC, and they can be
9038 used as a pair to hold a multiword value, then they @emph{must} have
9039 consecutive numbers after renumbering with @code{DBX_REGISTER_NUMBER}.
9040 Otherwise, debuggers will be unable to access such a pair, because they
9041 expect register pairs to be consecutive in their own numbering scheme.
9043 If you find yourself defining @code{DBX_REGISTER_NUMBER} in way that
9044 does not preserve register pairs, then what you must do instead is
9045 redefine the actual register numbering scheme.
9048 @defmac DEBUGGER_AUTO_OFFSET (@var{x})
9049 A C expression that returns the integer offset value for an automatic
9050 variable having address @var{x} (an RTL expression). The default
9051 computation assumes that @var{x} is based on the frame-pointer and
9052 gives the offset from the frame-pointer. This is required for targets
9053 that produce debugging output for DBX or COFF-style debugging output
9054 for SDB and allow the frame-pointer to be eliminated when the
9055 @option{-g} options is used.
9058 @defmac DEBUGGER_ARG_OFFSET (@var{offset}, @var{x})
9059 A C expression that returns the integer offset value for an argument
9060 having address @var{x} (an RTL expression). The nominal offset is
9064 @defmac PREFERRED_DEBUGGING_TYPE
9065 A C expression that returns the type of debugging output GCC should
9066 produce when the user specifies just @option{-g}. Define
9067 this if you have arranged for GCC to support more than one format of
9068 debugging output. Currently, the allowable values are @code{DBX_DEBUG},
9069 @code{SDB_DEBUG}, @code{DWARF_DEBUG}, @code{DWARF2_DEBUG},
9070 @code{XCOFF_DEBUG}, @code{VMS_DEBUG}, and @code{VMS_AND_DWARF2_DEBUG}.
9072 When the user specifies @option{-ggdb}, GCC normally also uses the
9073 value of this macro to select the debugging output format, but with two
9074 exceptions. If @code{DWARF2_DEBUGGING_INFO} is defined, GCC uses the
9075 value @code{DWARF2_DEBUG}. Otherwise, if @code{DBX_DEBUGGING_INFO} is
9076 defined, GCC uses @code{DBX_DEBUG}.
9078 The value of this macro only affects the default debugging output; the
9079 user can always get a specific type of output by using @option{-gstabs},
9080 @option{-gcoff}, @option{-gdwarf-2}, @option{-gxcoff}, or @option{-gvms}.
9084 @subsection Specific Options for DBX Output
9086 @c prevent bad page break with this line
9087 These are specific options for DBX output.
9089 @defmac DBX_DEBUGGING_INFO
9090 Define this macro if GCC should produce debugging output for DBX
9091 in response to the @option{-g} option.
9094 @defmac XCOFF_DEBUGGING_INFO
9095 Define this macro if GCC should produce XCOFF format debugging output
9096 in response to the @option{-g} option. This is a variant of DBX format.
9099 @defmac DEFAULT_GDB_EXTENSIONS
9100 Define this macro to control whether GCC should by default generate
9101 GDB's extended version of DBX debugging information (assuming DBX-format
9102 debugging information is enabled at all). If you don't define the
9103 macro, the default is 1: always generate the extended information
9104 if there is any occasion to.
9107 @defmac DEBUG_SYMS_TEXT
9108 Define this macro if all @code{.stabs} commands should be output while
9109 in the text section.
9112 @defmac ASM_STABS_OP
9113 A C string constant, including spacing, naming the assembler pseudo op to
9114 use instead of @code{"\t.stabs\t"} to define an ordinary debugging symbol.
9115 If you don't define this macro, @code{"\t.stabs\t"} is used. This macro
9116 applies only to DBX debugging information format.
9119 @defmac ASM_STABD_OP
9120 A C string constant, including spacing, naming the assembler pseudo op to
9121 use instead of @code{"\t.stabd\t"} to define a debugging symbol whose
9122 value is the current location. If you don't define this macro,
9123 @code{"\t.stabd\t"} is used. This macro applies only to DBX debugging
9127 @defmac ASM_STABN_OP
9128 A C string constant, including spacing, naming the assembler pseudo op to
9129 use instead of @code{"\t.stabn\t"} to define a debugging symbol with no
9130 name. If you don't define this macro, @code{"\t.stabn\t"} is used. This
9131 macro applies only to DBX debugging information format.
9134 @defmac DBX_NO_XREFS
9135 Define this macro if DBX on your system does not support the construct
9136 @samp{xs@var{tagname}}. On some systems, this construct is used to
9137 describe a forward reference to a structure named @var{tagname}.
9138 On other systems, this construct is not supported at all.
9141 @defmac DBX_CONTIN_LENGTH
9142 A symbol name in DBX-format debugging information is normally
9143 continued (split into two separate @code{.stabs} directives) when it
9144 exceeds a certain length (by default, 80 characters). On some
9145 operating systems, DBX requires this splitting; on others, splitting
9146 must not be done. You can inhibit splitting by defining this macro
9147 with the value zero. You can override the default splitting-length by
9148 defining this macro as an expression for the length you desire.
9151 @defmac DBX_CONTIN_CHAR
9152 Normally continuation is indicated by adding a @samp{\} character to
9153 the end of a @code{.stabs} string when a continuation follows. To use
9154 a different character instead, define this macro as a character
9155 constant for the character you want to use. Do not define this macro
9156 if backslash is correct for your system.
9159 @defmac DBX_STATIC_STAB_DATA_SECTION
9160 Define this macro if it is necessary to go to the data section before
9161 outputting the @samp{.stabs} pseudo-op for a non-global static
9165 @defmac DBX_TYPE_DECL_STABS_CODE
9166 The value to use in the ``code'' field of the @code{.stabs} directive
9167 for a typedef. The default is @code{N_LSYM}.
9170 @defmac DBX_STATIC_CONST_VAR_CODE
9171 The value to use in the ``code'' field of the @code{.stabs} directive
9172 for a static variable located in the text section. DBX format does not
9173 provide any ``right'' way to do this. The default is @code{N_FUN}.
9176 @defmac DBX_REGPARM_STABS_CODE
9177 The value to use in the ``code'' field of the @code{.stabs} directive
9178 for a parameter passed in registers. DBX format does not provide any
9179 ``right'' way to do this. The default is @code{N_RSYM}.
9182 @defmac DBX_REGPARM_STABS_LETTER
9183 The letter to use in DBX symbol data to identify a symbol as a parameter
9184 passed in registers. DBX format does not customarily provide any way to
9185 do this. The default is @code{'P'}.
9188 @defmac DBX_FUNCTION_FIRST
9189 Define this macro if the DBX information for a function and its
9190 arguments should precede the assembler code for the function. Normally,
9191 in DBX format, the debugging information entirely follows the assembler
9195 @defmac DBX_BLOCKS_FUNCTION_RELATIVE
9196 Define this macro, with value 1, if the value of a symbol describing
9197 the scope of a block (@code{N_LBRAC} or @code{N_RBRAC}) should be
9198 relative to the start of the enclosing function. Normally, GCC uses
9199 an absolute address.
9202 @defmac DBX_LINES_FUNCTION_RELATIVE
9203 Define this macro, with value 1, if the value of a symbol indicating
9204 the current line number (@code{N_SLINE}) should be relative to the
9205 start of the enclosing function. Normally, GCC uses an absolute address.
9208 @defmac DBX_USE_BINCL
9209 Define this macro if GCC should generate @code{N_BINCL} and
9210 @code{N_EINCL} stabs for included header files, as on Sun systems. This
9211 macro also directs GCC to output a type number as a pair of a file
9212 number and a type number within the file. Normally, GCC does not
9213 generate @code{N_BINCL} or @code{N_EINCL} stabs, and it outputs a single
9214 number for a type number.
9218 @subsection Open-Ended Hooks for DBX Format
9220 @c prevent bad page break with this line
9221 These are hooks for DBX format.
9223 @defmac DBX_OUTPUT_LBRAC (@var{stream}, @var{name})
9224 Define this macro to say how to output to @var{stream} the debugging
9225 information for the start of a scope level for variable names. The
9226 argument @var{name} is the name of an assembler symbol (for use with
9227 @code{assemble_name}) whose value is the address where the scope begins.
9230 @defmac DBX_OUTPUT_RBRAC (@var{stream}, @var{name})
9231 Like @code{DBX_OUTPUT_LBRAC}, but for the end of a scope level.
9234 @defmac DBX_OUTPUT_NFUN (@var{stream}, @var{lscope_label}, @var{decl})
9235 Define this macro if the target machine requires special handling to
9236 output an @code{N_FUN} entry for the function @var{decl}.
9239 @defmac DBX_OUTPUT_SOURCE_LINE (@var{stream}, @var{line}, @var{counter})
9240 A C statement to output DBX debugging information before code for line
9241 number @var{line} of the current source file to the stdio stream
9242 @var{stream}. @var{counter} is the number of time the macro was
9243 invoked, including the current invocation; it is intended to generate
9244 unique labels in the assembly output.
9246 This macro should not be defined if the default output is correct, or
9247 if it can be made correct by defining @code{DBX_LINES_FUNCTION_RELATIVE}.
9250 @defmac NO_DBX_FUNCTION_END
9251 Some stabs encapsulation formats (in particular ECOFF), cannot handle the
9252 @code{.stabs "",N_FUN,,0,0,Lscope-function-1} gdb dbx extension construct.
9253 On those machines, define this macro to turn this feature off without
9254 disturbing the rest of the gdb extensions.
9257 @defmac NO_DBX_BNSYM_ENSYM
9258 Some assemblers cannot handle the @code{.stabd BNSYM/ENSYM,0,0} gdb dbx
9259 extension construct. On those machines, define this macro to turn this
9260 feature off without disturbing the rest of the gdb extensions.
9263 @node File Names and DBX
9264 @subsection File Names in DBX Format
9266 @c prevent bad page break with this line
9267 This describes file names in DBX format.
9269 @defmac DBX_OUTPUT_MAIN_SOURCE_FILENAME (@var{stream}, @var{name})
9270 A C statement to output DBX debugging information to the stdio stream
9271 @var{stream}, which indicates that file @var{name} is the main source
9272 file---the file specified as the input file for compilation.
9273 This macro is called only once, at the beginning of compilation.
9275 This macro need not be defined if the standard form of output
9276 for DBX debugging information is appropriate.
9278 It may be necessary to refer to a label equal to the beginning of the
9279 text section. You can use @samp{assemble_name (stream, ltext_label_name)}
9280 to do so. If you do this, you must also set the variable
9281 @var{used_ltext_label_name} to @code{true}.
9284 @defmac NO_DBX_MAIN_SOURCE_DIRECTORY
9285 Define this macro, with value 1, if GCC should not emit an indication
9286 of the current directory for compilation and current source language at
9287 the beginning of the file.
9290 @defmac NO_DBX_GCC_MARKER
9291 Define this macro, with value 1, if GCC should not emit an indication
9292 that this object file was compiled by GCC@. The default is to emit
9293 an @code{N_OPT} stab at the beginning of every source file, with
9294 @samp{gcc2_compiled.} for the string and value 0.
9297 @defmac DBX_OUTPUT_MAIN_SOURCE_FILE_END (@var{stream}, @var{name})
9298 A C statement to output DBX debugging information at the end of
9299 compilation of the main source file @var{name}. Output should be
9300 written to the stdio stream @var{stream}.
9302 If you don't define this macro, nothing special is output at the end
9303 of compilation, which is correct for most machines.
9306 @defmac DBX_OUTPUT_NULL_N_SO_AT_MAIN_SOURCE_FILE_END
9307 Define this macro @emph{instead of} defining
9308 @code{DBX_OUTPUT_MAIN_SOURCE_FILE_END}, if what needs to be output at
9309 the end of compilation is an @code{N_SO} stab with an empty string,
9310 whose value is the highest absolute text address in the file.
9315 @subsection Macros for SDB and DWARF Output
9317 @c prevent bad page break with this line
9318 Here are macros for SDB and DWARF output.
9320 @defmac SDB_DEBUGGING_INFO
9321 Define this macro if GCC should produce COFF-style debugging output
9322 for SDB in response to the @option{-g} option.
9325 @defmac DWARF2_DEBUGGING_INFO
9326 Define this macro if GCC should produce dwarf version 2 format
9327 debugging output in response to the @option{-g} option.
9329 @hook TARGET_DWARF_CALLING_CONVENTION
9330 Define this to enable the dwarf attribute @code{DW_AT_calling_convention} to
9331 be emitted for each function. Instead of an integer return the enum
9332 value for the @code{DW_CC_} tag.
9335 To support optional call frame debugging information, you must also
9336 define @code{INCOMING_RETURN_ADDR_RTX} and either set
9337 @code{RTX_FRAME_RELATED_P} on the prologue insns if you use RTL for the
9338 prologue, or call @code{dwarf2out_def_cfa} and @code{dwarf2out_reg_save}
9339 as appropriate from @code{TARGET_ASM_FUNCTION_PROLOGUE} if you don't.
9342 @defmac DWARF2_FRAME_INFO
9343 Define this macro to a nonzero value if GCC should always output
9344 Dwarf 2 frame information. If @code{TARGET_EXCEPT_UNWIND_INFO}
9345 (@pxref{Exception Region Output}) returns @code{UI_DWARF2}, and
9346 exceptions are enabled, GCC will output this information not matter
9347 how you define @code{DWARF2_FRAME_INFO}.
9350 @hook TARGET_DEBUG_UNWIND_INFO
9351 This hook defines the mechanism that will be used for describing frame
9352 unwind information to the debugger. Normally the hook will return
9353 @code{UI_DWARF2} if DWARF 2 debug information is enabled, and
9354 return @code{UI_NONE} otherwise.
9356 A target may return @code{UI_DWARF2} even when DWARF 2 debug information
9357 is disabled in order to always output DWARF 2 frame information.
9359 A target may return @code{UI_TARGET} if it has ABI specified unwind tables.
9360 This will suppress generation of the normal debug frame unwind information.
9363 @defmac DWARF2_ASM_LINE_DEBUG_INFO
9364 Define this macro to be a nonzero value if the assembler can generate Dwarf 2
9365 line debug info sections. This will result in much more compact line number
9366 tables, and hence is desirable if it works.
9369 @hook TARGET_WANT_DEBUG_PUB_SECTIONS
9371 @hook TARGET_DELAY_SCHED2
9373 @hook TARGET_DELAY_VARTRACK
9375 @defmac ASM_OUTPUT_DWARF_DELTA (@var{stream}, @var{size}, @var{label1}, @var{label2})
9376 A C statement to issue assembly directives that create a difference
9377 @var{lab1} minus @var{lab2}, using an integer of the given @var{size}.
9380 @defmac ASM_OUTPUT_DWARF_VMS_DELTA (@var{stream}, @var{size}, @var{label1}, @var{label2})
9381 A C statement to issue assembly directives that create a difference
9382 between the two given labels in system defined units, e.g. instruction
9383 slots on IA64 VMS, using an integer of the given size.
9386 @defmac ASM_OUTPUT_DWARF_OFFSET (@var{stream}, @var{size}, @var{label}, @var{section})
9387 A C statement to issue assembly directives that create a
9388 section-relative reference to the given @var{label}, using an integer of the
9389 given @var{size}. The label is known to be defined in the given @var{section}.
9392 @defmac ASM_OUTPUT_DWARF_PCREL (@var{stream}, @var{size}, @var{label})
9393 A C statement to issue assembly directives that create a self-relative
9394 reference to the given @var{label}, using an integer of the given @var{size}.
9397 @defmac ASM_OUTPUT_DWARF_TABLE_REF (@var{label})
9398 A C statement to issue assembly directives that create a reference to
9399 the DWARF table identifier @var{label} from the current section. This
9400 is used on some systems to avoid garbage collecting a DWARF table which
9401 is referenced by a function.
9404 @hook TARGET_ASM_OUTPUT_DWARF_DTPREL
9405 If defined, this target hook is a function which outputs a DTP-relative
9406 reference to the given TLS symbol of the specified size.
9409 @defmac PUT_SDB_@dots{}
9410 Define these macros to override the assembler syntax for the special
9411 SDB assembler directives. See @file{sdbout.c} for a list of these
9412 macros and their arguments. If the standard syntax is used, you need
9413 not define them yourself.
9417 Some assemblers do not support a semicolon as a delimiter, even between
9418 SDB assembler directives. In that case, define this macro to be the
9419 delimiter to use (usually @samp{\n}). It is not necessary to define
9420 a new set of @code{PUT_SDB_@var{op}} macros if this is the only change
9424 @defmac SDB_ALLOW_UNKNOWN_REFERENCES
9425 Define this macro to allow references to unknown structure,
9426 union, or enumeration tags to be emitted. Standard COFF does not
9427 allow handling of unknown references, MIPS ECOFF has support for
9431 @defmac SDB_ALLOW_FORWARD_REFERENCES
9432 Define this macro to allow references to structure, union, or
9433 enumeration tags that have not yet been seen to be handled. Some
9434 assemblers choke if forward tags are used, while some require it.
9437 @defmac SDB_OUTPUT_SOURCE_LINE (@var{stream}, @var{line})
9438 A C statement to output SDB debugging information before code for line
9439 number @var{line} of the current source file to the stdio stream
9440 @var{stream}. The default is to emit an @code{.ln} directive.
9445 @subsection Macros for VMS Debug Format
9447 @c prevent bad page break with this line
9448 Here are macros for VMS debug format.
9450 @defmac VMS_DEBUGGING_INFO
9451 Define this macro if GCC should produce debugging output for VMS
9452 in response to the @option{-g} option. The default behavior for VMS
9453 is to generate minimal debug info for a traceback in the absence of
9454 @option{-g} unless explicitly overridden with @option{-g0}. This
9455 behavior is controlled by @code{TARGET_OPTION_OPTIMIZATION} and
9456 @code{TARGET_OPTION_OVERRIDE}.
9459 @node Floating Point
9460 @section Cross Compilation and Floating Point
9461 @cindex cross compilation and floating point
9462 @cindex floating point and cross compilation
9464 While all modern machines use twos-complement representation for integers,
9465 there are a variety of representations for floating point numbers. This
9466 means that in a cross-compiler the representation of floating point numbers
9467 in the compiled program may be different from that used in the machine
9468 doing the compilation.
9470 Because different representation systems may offer different amounts of
9471 range and precision, all floating point constants must be represented in
9472 the target machine's format. Therefore, the cross compiler cannot
9473 safely use the host machine's floating point arithmetic; it must emulate
9474 the target's arithmetic. To ensure consistency, GCC always uses
9475 emulation to work with floating point values, even when the host and
9476 target floating point formats are identical.
9478 The following macros are provided by @file{real.h} for the compiler to
9479 use. All parts of the compiler which generate or optimize
9480 floating-point calculations must use these macros. They may evaluate
9481 their operands more than once, so operands must not have side effects.
9483 @defmac REAL_VALUE_TYPE
9484 The C data type to be used to hold a floating point value in the target
9485 machine's format. Typically this is a @code{struct} containing an
9486 array of @code{HOST_WIDE_INT}, but all code should treat it as an opaque
9490 @deftypefn Macro int REAL_VALUES_EQUAL (REAL_VALUE_TYPE @var{x}, REAL_VALUE_TYPE @var{y})
9491 Compares for equality the two values, @var{x} and @var{y}. If the target
9492 floating point format supports negative zeroes and/or NaNs,
9493 @samp{REAL_VALUES_EQUAL (-0.0, 0.0)} is true, and
9494 @samp{REAL_VALUES_EQUAL (NaN, NaN)} is false.
9497 @deftypefn Macro int REAL_VALUES_LESS (REAL_VALUE_TYPE @var{x}, REAL_VALUE_TYPE @var{y})
9498 Tests whether @var{x} is less than @var{y}.
9501 @deftypefn Macro HOST_WIDE_INT REAL_VALUE_FIX (REAL_VALUE_TYPE @var{x})
9502 Truncates @var{x} to a signed integer, rounding toward zero.
9505 @deftypefn Macro {unsigned HOST_WIDE_INT} REAL_VALUE_UNSIGNED_FIX (REAL_VALUE_TYPE @var{x})
9506 Truncates @var{x} to an unsigned integer, rounding toward zero. If
9507 @var{x} is negative, returns zero.
9510 @deftypefn Macro REAL_VALUE_TYPE REAL_VALUE_ATOF (const char *@var{string}, enum machine_mode @var{mode})
9511 Converts @var{string} into a floating point number in the target machine's
9512 representation for mode @var{mode}. This routine can handle both
9513 decimal and hexadecimal floating point constants, using the syntax
9514 defined by the C language for both.
9517 @deftypefn Macro int REAL_VALUE_NEGATIVE (REAL_VALUE_TYPE @var{x})
9518 Returns 1 if @var{x} is negative (including negative zero), 0 otherwise.
9521 @deftypefn Macro int REAL_VALUE_ISINF (REAL_VALUE_TYPE @var{x})
9522 Determines whether @var{x} represents infinity (positive or negative).
9525 @deftypefn Macro int REAL_VALUE_ISNAN (REAL_VALUE_TYPE @var{x})
9526 Determines whether @var{x} represents a ``NaN'' (not-a-number).
9529 @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})
9530 Calculates an arithmetic operation on the two floating point values
9531 @var{x} and @var{y}, storing the result in @var{output} (which must be a
9534 The operation to be performed is specified by @var{code}. Only the
9535 following codes are supported: @code{PLUS_EXPR}, @code{MINUS_EXPR},
9536 @code{MULT_EXPR}, @code{RDIV_EXPR}, @code{MAX_EXPR}, @code{MIN_EXPR}.
9538 If @code{REAL_ARITHMETIC} is asked to evaluate division by zero and the
9539 target's floating point format cannot represent infinity, it will call
9540 @code{abort}. Callers should check for this situation first, using
9541 @code{MODE_HAS_INFINITIES}. @xref{Storage Layout}.
9544 @deftypefn Macro REAL_VALUE_TYPE REAL_VALUE_NEGATE (REAL_VALUE_TYPE @var{x})
9545 Returns the negative of the floating point value @var{x}.
9548 @deftypefn Macro REAL_VALUE_TYPE REAL_VALUE_ABS (REAL_VALUE_TYPE @var{x})
9549 Returns the absolute value of @var{x}.
9552 @deftypefn Macro REAL_VALUE_TYPE REAL_VALUE_TRUNCATE (REAL_VALUE_TYPE @var{mode}, enum machine_mode @var{x})
9553 Truncates the floating point value @var{x} to fit in @var{mode}. The
9554 return value is still a full-size @code{REAL_VALUE_TYPE}, but it has an
9555 appropriate bit pattern to be output as a floating constant whose
9556 precision accords with mode @var{mode}.
9559 @deftypefn Macro void REAL_VALUE_TO_INT (HOST_WIDE_INT @var{low}, HOST_WIDE_INT @var{high}, REAL_VALUE_TYPE @var{x})
9560 Converts a floating point value @var{x} into a double-precision integer
9561 which is then stored into @var{low} and @var{high}. If the value is not
9562 integral, it is truncated.
9565 @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})
9566 Converts a double-precision integer found in @var{low} and @var{high},
9567 into a floating point value which is then stored into @var{x}. The
9568 value is truncated to fit in mode @var{mode}.
9571 @node Mode Switching
9572 @section Mode Switching Instructions
9573 @cindex mode switching
9574 The following macros control mode switching optimizations:
9576 @defmac OPTIMIZE_MODE_SWITCHING (@var{entity})
9577 Define this macro if the port needs extra instructions inserted for mode
9578 switching in an optimizing compilation.
9580 For an example, the SH4 can perform both single and double precision
9581 floating point operations, but to perform a single precision operation,
9582 the FPSCR PR bit has to be cleared, while for a double precision
9583 operation, this bit has to be set. Changing the PR bit requires a general
9584 purpose register as a scratch register, hence these FPSCR sets have to
9585 be inserted before reload, i.e.@: you can't put this into instruction emitting
9586 or @code{TARGET_MACHINE_DEPENDENT_REORG}.
9588 You can have multiple entities that are mode-switched, and select at run time
9589 which entities actually need it. @code{OPTIMIZE_MODE_SWITCHING} should
9590 return nonzero for any @var{entity} that needs mode-switching.
9591 If you define this macro, you also have to define
9592 @code{NUM_MODES_FOR_MODE_SWITCHING}, @code{MODE_NEEDED},
9593 @code{MODE_PRIORITY_TO_MODE} and @code{EMIT_MODE_SET}.
9594 @code{MODE_AFTER}, @code{MODE_ENTRY}, and @code{MODE_EXIT}
9598 @defmac NUM_MODES_FOR_MODE_SWITCHING
9599 If you define @code{OPTIMIZE_MODE_SWITCHING}, you have to define this as
9600 initializer for an array of integers. Each initializer element
9601 N refers to an entity that needs mode switching, and specifies the number
9602 of different modes that might need to be set for this entity.
9603 The position of the initializer in the initializer---starting counting at
9604 zero---determines the integer that is used to refer to the mode-switched
9606 In macros that take mode arguments / yield a mode result, modes are
9607 represented as numbers 0 @dots{} N @minus{} 1. N is used to specify that no mode
9608 switch is needed / supplied.
9611 @defmac MODE_NEEDED (@var{entity}, @var{insn})
9612 @var{entity} is an integer specifying a mode-switched entity. If
9613 @code{OPTIMIZE_MODE_SWITCHING} is defined, you must define this macro to
9614 return an integer value not larger than the corresponding element in
9615 @code{NUM_MODES_FOR_MODE_SWITCHING}, to denote the mode that @var{entity} must
9616 be switched into prior to the execution of @var{insn}.
9619 @defmac MODE_AFTER (@var{mode}, @var{insn})
9620 If this macro is defined, it is evaluated for every @var{insn} during
9621 mode switching. It determines the mode that an insn results in (if
9622 different from the incoming mode).
9625 @defmac MODE_ENTRY (@var{entity})
9626 If this macro is defined, it is evaluated for every @var{entity} that needs
9627 mode switching. It should evaluate to an integer, which is a mode that
9628 @var{entity} is assumed to be switched to at function entry. If @code{MODE_ENTRY}
9629 is defined then @code{MODE_EXIT} must be defined.
9632 @defmac MODE_EXIT (@var{entity})
9633 If this macro is defined, it is evaluated for every @var{entity} that needs
9634 mode switching. It should evaluate to an integer, which is a mode that
9635 @var{entity} is assumed to be switched to at function exit. If @code{MODE_EXIT}
9636 is defined then @code{MODE_ENTRY} must be defined.
9639 @defmac MODE_PRIORITY_TO_MODE (@var{entity}, @var{n})
9640 This macro specifies the order in which modes for @var{entity} are processed.
9641 0 is the highest priority, @code{NUM_MODES_FOR_MODE_SWITCHING[@var{entity}] - 1} the
9642 lowest. The value of the macro should be an integer designating a mode
9643 for @var{entity}. For any fixed @var{entity}, @code{mode_priority_to_mode}
9644 (@var{entity}, @var{n}) shall be a bijection in 0 @dots{}
9645 @code{num_modes_for_mode_switching[@var{entity}] - 1}.
9648 @defmac EMIT_MODE_SET (@var{entity}, @var{mode}, @var{hard_regs_live})
9649 Generate one or more insns to set @var{entity} to @var{mode}.
9650 @var{hard_reg_live} is the set of hard registers live at the point where
9651 the insn(s) are to be inserted.
9654 @node Target Attributes
9655 @section Defining target-specific uses of @code{__attribute__}
9656 @cindex target attributes
9657 @cindex machine attributes
9658 @cindex attributes, target-specific
9660 Target-specific attributes may be defined for functions, data and types.
9661 These are described using the following target hooks; they also need to
9662 be documented in @file{extend.texi}.
9664 @hook TARGET_ATTRIBUTE_TABLE
9665 If defined, this target hook points to an array of @samp{struct
9666 attribute_spec} (defined in @file{tree.h}) specifying the machine
9667 specific attributes for this target and some of the restrictions on the
9668 entities to which these attributes are applied and the arguments they
9672 @hook TARGET_ATTRIBUTE_TAKES_IDENTIFIER_P
9673 If defined, this target hook is a function which returns true if the
9674 machine-specific attribute named @var{name} expects an identifier
9675 given as its first argument to be passed on as a plain identifier, not
9676 subjected to name lookup. If this is not defined, the default is
9677 false for all machine-specific attributes.
9680 @hook TARGET_COMP_TYPE_ATTRIBUTES
9681 If defined, this target hook is a function which returns zero if the attributes on
9682 @var{type1} and @var{type2} are incompatible, one if they are compatible,
9683 and two if they are nearly compatible (which causes a warning to be
9684 generated). If this is not defined, machine-specific attributes are
9685 supposed always to be compatible.
9688 @hook TARGET_SET_DEFAULT_TYPE_ATTRIBUTES
9689 If defined, this target hook is a function which assigns default attributes to
9690 the newly defined @var{type}.
9693 @hook TARGET_MERGE_TYPE_ATTRIBUTES
9694 Define this target hook if the merging of type attributes needs special
9695 handling. If defined, the result is a list of the combined
9696 @code{TYPE_ATTRIBUTES} of @var{type1} and @var{type2}. It is assumed
9697 that @code{comptypes} has already been called and returned 1. This
9698 function may call @code{merge_attributes} to handle machine-independent
9702 @hook TARGET_MERGE_DECL_ATTRIBUTES
9703 Define this target hook if the merging of decl attributes needs special
9704 handling. If defined, the result is a list of the combined
9705 @code{DECL_ATTRIBUTES} of @var{olddecl} and @var{newdecl}.
9706 @var{newdecl} is a duplicate declaration of @var{olddecl}. Examples of
9707 when this is needed are when one attribute overrides another, or when an
9708 attribute is nullified by a subsequent definition. This function may
9709 call @code{merge_attributes} to handle machine-independent merging.
9711 @findex TARGET_DLLIMPORT_DECL_ATTRIBUTES
9712 If the only target-specific handling you require is @samp{dllimport}
9713 for Microsoft Windows targets, you should define the macro
9714 @code{TARGET_DLLIMPORT_DECL_ATTRIBUTES} to @code{1}. The compiler
9715 will then define a function called
9716 @code{merge_dllimport_decl_attributes} which can then be defined as
9717 the expansion of @code{TARGET_MERGE_DECL_ATTRIBUTES}. You can also
9718 add @code{handle_dll_attribute} in the attribute table for your port
9719 to perform initial processing of the @samp{dllimport} and
9720 @samp{dllexport} attributes. This is done in @file{i386/cygwin.h} and
9721 @file{i386/i386.c}, for example.
9724 @hook TARGET_VALID_DLLIMPORT_ATTRIBUTE_P
9726 @defmac TARGET_DECLSPEC
9727 Define this macro to a nonzero value if you want to treat
9728 @code{__declspec(X)} as equivalent to @code{__attribute((X))}. By
9729 default, this behavior is enabled only for targets that define
9730 @code{TARGET_DLLIMPORT_DECL_ATTRIBUTES}. The current implementation
9731 of @code{__declspec} is via a built-in macro, but you should not rely
9732 on this implementation detail.
9735 @hook TARGET_INSERT_ATTRIBUTES
9736 Define this target hook if you want to be able to add attributes to a decl
9737 when it is being created. This is normally useful for back ends which
9738 wish to implement a pragma by using the attributes which correspond to
9739 the pragma's effect. The @var{node} argument is the decl which is being
9740 created. The @var{attr_ptr} argument is a pointer to the attribute list
9741 for this decl. The list itself should not be modified, since it may be
9742 shared with other decls, but attributes may be chained on the head of
9743 the list and @code{*@var{attr_ptr}} modified to point to the new
9744 attributes, or a copy of the list may be made if further changes are
9748 @hook TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P
9750 This target hook returns @code{true} if it is ok to inline @var{fndecl}
9751 into the current function, despite its having target-specific
9752 attributes, @code{false} otherwise. By default, if a function has a
9753 target specific attribute attached to it, it will not be inlined.
9756 @hook TARGET_OPTION_VALID_ATTRIBUTE_P
9757 This hook is called to parse the @code{attribute(option("..."))}, and
9758 it allows the function to set different target machine compile time
9759 options for the current function that might be different than the
9760 options specified on the command line. The hook should return
9761 @code{true} if the options are valid.
9763 The hook should set the @var{DECL_FUNCTION_SPECIFIC_TARGET} field in
9764 the function declaration to hold a pointer to a target specific
9765 @var{struct cl_target_option} structure.
9768 @hook TARGET_OPTION_SAVE
9769 This hook is called to save any additional target specific information
9770 in the @var{struct cl_target_option} structure for function specific
9772 @xref{Option file format}.
9775 @hook TARGET_OPTION_RESTORE
9776 This hook is called to restore any additional target specific
9777 information in the @var{struct cl_target_option} structure for
9778 function specific options.
9781 @hook TARGET_OPTION_PRINT
9782 This hook is called to print any additional target specific
9783 information in the @var{struct cl_target_option} structure for
9784 function specific options.
9787 @hook TARGET_OPTION_PRAGMA_PARSE
9788 This target hook parses the options for @code{#pragma GCC option} to
9789 set the machine specific options for functions that occur later in the
9790 input stream. The options should be the same as handled by the
9791 @code{TARGET_OPTION_VALID_ATTRIBUTE_P} hook.
9794 @hook TARGET_OPTION_OVERRIDE
9795 Sometimes certain combinations of command options do not make sense on
9796 a particular target machine. You can override the hook
9797 @code{TARGET_OPTION_OVERRIDE} to take account of this. This hooks is called
9798 once just after all the command options have been parsed.
9800 Don't use this hook to turn on various extra optimizations for
9801 @option{-O}. That is what @code{TARGET_OPTION_OPTIMIZATION} is for.
9803 If you need to do something whenever the optimization level is
9804 changed via the optimize attribute or pragma, see
9805 @code{TARGET_OVERRIDE_OPTIONS_AFTER_CHANGE}
9808 @hook TARGET_CAN_INLINE_P
9809 This target hook returns @code{false} if the @var{caller} function
9810 cannot inline @var{callee}, based on target specific information. By
9811 default, inlining is not allowed if the callee function has function
9812 specific target options and the caller does not use the same options.
9816 @section Emulating TLS
9817 @cindex Emulated TLS
9819 For targets whose psABI does not provide Thread Local Storage via
9820 specific relocations and instruction sequences, an emulation layer is
9821 used. A set of target hooks allows this emulation layer to be
9822 configured for the requirements of a particular target. For instance
9823 the psABI may in fact specify TLS support in terms of an emulation
9826 The emulation layer works by creating a control object for every TLS
9827 object. To access the TLS object, a lookup function is provided
9828 which, when given the address of the control object, will return the
9829 address of the current thread's instance of the TLS object.
9831 @hook TARGET_EMUTLS_GET_ADDRESS
9832 Contains the name of the helper function that uses a TLS control
9833 object to locate a TLS instance. The default causes libgcc's
9834 emulated TLS helper function to be used.
9837 @hook TARGET_EMUTLS_REGISTER_COMMON
9838 Contains the name of the helper function that should be used at
9839 program startup to register TLS objects that are implicitly
9840 initialized to zero. If this is @code{NULL}, all TLS objects will
9841 have explicit initializers. The default causes libgcc's emulated TLS
9842 registration function to be used.
9845 @hook TARGET_EMUTLS_VAR_SECTION
9846 Contains the name of the section in which TLS control variables should
9847 be placed. The default of @code{NULL} allows these to be placed in
9851 @hook TARGET_EMUTLS_TMPL_SECTION
9852 Contains the name of the section in which TLS initializers should be
9853 placed. The default of @code{NULL} allows these to be placed in any
9857 @hook TARGET_EMUTLS_VAR_PREFIX
9858 Contains the prefix to be prepended to TLS control variable names.
9859 The default of @code{NULL} uses a target-specific prefix.
9862 @hook TARGET_EMUTLS_TMPL_PREFIX
9863 Contains the prefix to be prepended to TLS initializer objects. The
9864 default of @code{NULL} uses a target-specific prefix.
9867 @hook TARGET_EMUTLS_VAR_FIELDS
9868 Specifies a function that generates the FIELD_DECLs for a TLS control
9869 object type. @var{type} is the RECORD_TYPE the fields are for and
9870 @var{name} should be filled with the structure tag, if the default of
9871 @code{__emutls_object} is unsuitable. The default creates a type suitable
9872 for libgcc's emulated TLS function.
9875 @hook TARGET_EMUTLS_VAR_INIT
9876 Specifies a function that generates the CONSTRUCTOR to initialize a
9877 TLS control object. @var{var} is the TLS control object, @var{decl}
9878 is the TLS object and @var{tmpl_addr} is the address of the
9879 initializer. The default initializes libgcc's emulated TLS control object.
9882 @hook TARGET_EMUTLS_VAR_ALIGN_FIXED
9883 Specifies whether the alignment of TLS control variable objects is
9884 fixed and should not be increased as some backends may do to optimize
9885 single objects. The default is false.
9888 @hook TARGET_EMUTLS_DEBUG_FORM_TLS_ADDRESS
9889 Specifies whether a DWARF @code{DW_OP_form_tls_address} location descriptor
9890 may be used to describe emulated TLS control objects.
9893 @node MIPS Coprocessors
9894 @section Defining coprocessor specifics for MIPS targets.
9895 @cindex MIPS coprocessor-definition macros
9897 The MIPS specification allows MIPS implementations to have as many as 4
9898 coprocessors, each with as many as 32 private registers. GCC supports
9899 accessing these registers and transferring values between the registers
9900 and memory using asm-ized variables. For example:
9903 register unsigned int cp0count asm ("c0r1");
9909 (``c0r1'' is the default name of register 1 in coprocessor 0; alternate
9910 names may be added as described below, or the default names may be
9911 overridden entirely in @code{SUBTARGET_CONDITIONAL_REGISTER_USAGE}.)
9913 Coprocessor registers are assumed to be epilogue-used; sets to them will
9914 be preserved even if it does not appear that the register is used again
9915 later in the function.
9917 Another note: according to the MIPS spec, coprocessor 1 (if present) is
9918 the FPU@. One accesses COP1 registers through standard mips
9919 floating-point support; they are not included in this mechanism.
9921 There is one macro used in defining the MIPS coprocessor interface which
9922 you may want to override in subtargets; it is described below.
9924 @defmac ALL_COP_ADDITIONAL_REGISTER_NAMES
9925 A comma-separated list (with leading comma) of pairs describing the
9926 alternate names of coprocessor registers. The format of each entry should be
9928 @{ @var{alternatename}, @var{register_number}@}
9934 @section Parameters for Precompiled Header Validity Checking
9935 @cindex parameters, precompiled headers
9937 @hook TARGET_GET_PCH_VALIDITY
9938 This hook returns a pointer to the data needed by
9939 @code{TARGET_PCH_VALID_P} and sets
9940 @samp{*@var{sz}} to the size of the data in bytes.
9943 @hook TARGET_PCH_VALID_P
9944 This hook checks whether the options used to create a PCH file are
9945 compatible with the current settings. It returns @code{NULL}
9946 if so and a suitable error message if not. Error messages will
9947 be presented to the user and must be localized using @samp{_(@var{msg})}.
9949 @var{data} is the data that was returned by @code{TARGET_GET_PCH_VALIDITY}
9950 when the PCH file was created and @var{sz} is the size of that data in bytes.
9951 It's safe to assume that the data was created by the same version of the
9952 compiler, so no format checking is needed.
9954 The default definition of @code{default_pch_valid_p} should be
9955 suitable for most targets.
9958 @hook TARGET_CHECK_PCH_TARGET_FLAGS
9959 If this hook is nonnull, the default implementation of
9960 @code{TARGET_PCH_VALID_P} will use it to check for compatible values
9961 of @code{target_flags}. @var{pch_flags} specifies the value that
9962 @code{target_flags} had when the PCH file was created. The return
9963 value is the same as for @code{TARGET_PCH_VALID_P}.
9967 @section C++ ABI parameters
9968 @cindex parameters, c++ abi
9970 @hook TARGET_CXX_GUARD_TYPE
9971 Define this hook to override the integer type used for guard variables.
9972 These are used to implement one-time construction of static objects. The
9973 default is long_long_integer_type_node.
9976 @hook TARGET_CXX_GUARD_MASK_BIT
9977 This hook determines how guard variables are used. It should return
9978 @code{false} (the default) if the first byte should be used. A return value of
9979 @code{true} indicates that only the least significant bit should be used.
9982 @hook TARGET_CXX_GET_COOKIE_SIZE
9983 This hook returns the size of the cookie to use when allocating an array
9984 whose elements have the indicated @var{type}. Assumes that it is already
9985 known that a cookie is needed. The default is
9986 @code{max(sizeof (size_t), alignof(type))}, as defined in section 2.7 of the
9987 IA64/Generic C++ ABI@.
9990 @hook TARGET_CXX_COOKIE_HAS_SIZE
9991 This hook should return @code{true} if the element size should be stored in
9992 array cookies. The default is to return @code{false}.
9995 @hook TARGET_CXX_IMPORT_EXPORT_CLASS
9996 If defined by a backend this hook allows the decision made to export
9997 class @var{type} to be overruled. Upon entry @var{import_export}
9998 will contain 1 if the class is going to be exported, @minus{}1 if it is going
9999 to be imported and 0 otherwise. This function should return the
10000 modified value and perform any other actions necessary to support the
10001 backend's targeted operating system.
10004 @hook TARGET_CXX_CDTOR_RETURNS_THIS
10005 This hook should return @code{true} if constructors and destructors return
10006 the address of the object created/destroyed. The default is to return
10010 @hook TARGET_CXX_KEY_METHOD_MAY_BE_INLINE
10011 This hook returns true if the key method for a class (i.e., the method
10012 which, if defined in the current translation unit, causes the virtual
10013 table to be emitted) may be an inline function. Under the standard
10014 Itanium C++ ABI the key method may be an inline function so long as
10015 the function is not declared inline in the class definition. Under
10016 some variants of the ABI, an inline function can never be the key
10017 method. The default is to return @code{true}.
10020 @hook TARGET_CXX_DETERMINE_CLASS_DATA_VISIBILITY
10022 @hook TARGET_CXX_CLASS_DATA_ALWAYS_COMDAT
10023 This hook returns true (the default) if virtual tables and other
10024 similar implicit class data objects are always COMDAT if they have
10025 external linkage. If this hook returns false, then class data for
10026 classes whose virtual table will be emitted in only one translation
10027 unit will not be COMDAT.
10030 @hook TARGET_CXX_LIBRARY_RTTI_COMDAT
10031 This hook returns true (the default) if the RTTI information for
10032 the basic types which is defined in the C++ runtime should always
10033 be COMDAT, false if it should not be COMDAT.
10036 @hook TARGET_CXX_USE_AEABI_ATEXIT
10037 This hook returns true if @code{__aeabi_atexit} (as defined by the ARM EABI)
10038 should be used to register static destructors when @option{-fuse-cxa-atexit}
10039 is in effect. The default is to return false to use @code{__cxa_atexit}.
10042 @hook TARGET_CXX_USE_ATEXIT_FOR_CXA_ATEXIT
10043 This hook returns true if the target @code{atexit} function can be used
10044 in the same manner as @code{__cxa_atexit} to register C++ static
10045 destructors. This requires that @code{atexit}-registered functions in
10046 shared libraries are run in the correct order when the libraries are
10047 unloaded. The default is to return false.
10050 @hook TARGET_CXX_ADJUST_CLASS_AT_DEFINITION
10052 @node Named Address Spaces
10053 @section Adding support for named address spaces
10054 @cindex named address spaces
10056 The draft technical report of the ISO/IEC JTC1 S22 WG14 N1275
10057 standards committee, @cite{Programming Languages - C - Extensions to
10058 support embedded processors}, specifies a syntax for embedded
10059 processors to specify alternate address spaces. You can configure a
10060 GCC port to support section 5.1 of the draft report to add support for
10061 address spaces other than the default address space. These address
10062 spaces are new keywords that are similar to the @code{volatile} and
10063 @code{const} type attributes.
10065 Pointers to named address spaces can have a different size than
10066 pointers to the generic address space.
10068 For example, the SPU port uses the @code{__ea} address space to refer
10069 to memory in the host processor, rather than memory local to the SPU
10070 processor. Access to memory in the @code{__ea} address space involves
10071 issuing DMA operations to move data between the host processor and the
10072 local processor memory address space. Pointers in the @code{__ea}
10073 address space are either 32 bits or 64 bits based on the
10074 @option{-mea32} or @option{-mea64} switches (native SPU pointers are
10077 Internally, address spaces are represented as a small integer in the
10078 range 0 to 15 with address space 0 being reserved for the generic
10081 To register a named address space qualifier keyword with the C front end,
10082 the target may call the @code{c_register_addr_space} routine. For example,
10083 the SPU port uses the following to declare @code{__ea} as the keyword for
10084 named address space #1:
10086 #define ADDR_SPACE_EA 1
10087 c_register_addr_space ("__ea", ADDR_SPACE_EA);
10090 @hook TARGET_ADDR_SPACE_POINTER_MODE
10091 Define this to return the machine mode to use for pointers to
10092 @var{address_space} if the target supports named address spaces.
10093 The default version of this hook returns @code{ptr_mode} for the
10094 generic address space only.
10097 @hook TARGET_ADDR_SPACE_ADDRESS_MODE
10098 Define this to return the machine mode to use for addresses in
10099 @var{address_space} if the target supports named address spaces.
10100 The default version of this hook returns @code{Pmode} for the
10101 generic address space only.
10104 @hook TARGET_ADDR_SPACE_VALID_POINTER_MODE
10105 Define this to return nonzero if the port can handle pointers
10106 with machine mode @var{mode} to address space @var{as}. This target
10107 hook is the same as the @code{TARGET_VALID_POINTER_MODE} target hook,
10108 except that it includes explicit named address space support. The default
10109 version of this hook returns true for the modes returned by either the
10110 @code{TARGET_ADDR_SPACE_POINTER_MODE} or @code{TARGET_ADDR_SPACE_ADDRESS_MODE}
10111 target hooks for the given address space.
10114 @hook TARGET_ADDR_SPACE_LEGITIMATE_ADDRESS_P
10115 Define this to return true if @var{exp} is a valid address for mode
10116 @var{mode} in the named address space @var{as}. The @var{strict}
10117 parameter says whether strict addressing is in effect after reload has
10118 finished. This target hook is the same as the
10119 @code{TARGET_LEGITIMATE_ADDRESS_P} target hook, except that it includes
10120 explicit named address space support.
10123 @hook TARGET_ADDR_SPACE_LEGITIMIZE_ADDRESS
10124 Define this to modify an invalid address @var{x} to be a valid address
10125 with mode @var{mode} in the named address space @var{as}. This target
10126 hook is the same as the @code{TARGET_LEGITIMIZE_ADDRESS} target hook,
10127 except that it includes explicit named address space support.
10130 @hook TARGET_ADDR_SPACE_SUBSET_P
10131 Define this to return whether the @var{subset} named address space is
10132 contained within the @var{superset} named address space. Pointers to
10133 a named address space that is a subset of another named address space
10134 will be converted automatically without a cast if used together in
10135 arithmetic operations. Pointers to a superset address space can be
10136 converted to pointers to a subset address space via explicit casts.
10139 @hook TARGET_ADDR_SPACE_CONVERT
10140 Define this to convert the pointer expression represented by the RTL
10141 @var{op} with type @var{from_type} that points to a named address
10142 space to a new pointer expression with type @var{to_type} that points
10143 to a different named address space. When this hook it called, it is
10144 guaranteed that one of the two address spaces is a subset of the other,
10145 as determined by the @code{TARGET_ADDR_SPACE_SUBSET_P} target hook.
10149 @section Miscellaneous Parameters
10150 @cindex parameters, miscellaneous
10152 @c prevent bad page break with this line
10153 Here are several miscellaneous parameters.
10155 @defmac HAS_LONG_COND_BRANCH
10156 Define this boolean macro to indicate whether or not your architecture
10157 has conditional branches that can span all of memory. It is used in
10158 conjunction with an optimization that partitions hot and cold basic
10159 blocks into separate sections of the executable. If this macro is
10160 set to false, gcc will convert any conditional branches that attempt
10161 to cross between sections into unconditional branches or indirect jumps.
10164 @defmac HAS_LONG_UNCOND_BRANCH
10165 Define this boolean macro to indicate whether or not your architecture
10166 has unconditional branches that can span all of memory. It is used in
10167 conjunction with an optimization that partitions hot and cold basic
10168 blocks into separate sections of the executable. If this macro is
10169 set to false, gcc will convert any unconditional branches that attempt
10170 to cross between sections into indirect jumps.
10173 @defmac CASE_VECTOR_MODE
10174 An alias for a machine mode name. This is the machine mode that
10175 elements of a jump-table should have.
10178 @defmac CASE_VECTOR_SHORTEN_MODE (@var{min_offset}, @var{max_offset}, @var{body})
10179 Optional: return the preferred mode for an @code{addr_diff_vec}
10180 when the minimum and maximum offset are known. If you define this,
10181 it enables extra code in branch shortening to deal with @code{addr_diff_vec}.
10182 To make this work, you also have to define @code{INSN_ALIGN} and
10183 make the alignment for @code{addr_diff_vec} explicit.
10184 The @var{body} argument is provided so that the offset_unsigned and scale
10185 flags can be updated.
10188 @defmac CASE_VECTOR_PC_RELATIVE
10189 Define this macro to be a C expression to indicate when jump-tables
10190 should contain relative addresses. You need not define this macro if
10191 jump-tables never contain relative addresses, or jump-tables should
10192 contain relative addresses only when @option{-fPIC} or @option{-fPIC}
10196 @hook TARGET_CASE_VALUES_THRESHOLD
10197 This function return the smallest number of different values for which it
10198 is best to use a jump-table instead of a tree of conditional branches.
10199 The default is four for machines with a @code{casesi} instruction and
10200 five otherwise. This is best for most machines.
10203 @defmac CASE_USE_BIT_TESTS
10204 Define this macro to be a C expression to indicate whether C switch
10205 statements may be implemented by a sequence of bit tests. This is
10206 advantageous on processors that can efficiently implement left shift
10207 of 1 by the number of bits held in a register, but inappropriate on
10208 targets that would require a loop. By default, this macro returns
10209 @code{true} if the target defines an @code{ashlsi3} pattern, and
10210 @code{false} otherwise.
10213 @defmac WORD_REGISTER_OPERATIONS
10214 Define this macro if operations between registers with integral mode
10215 smaller than a word are always performed on the entire register.
10216 Most RISC machines have this property and most CISC machines do not.
10219 @defmac LOAD_EXTEND_OP (@var{mem_mode})
10220 Define this macro to be a C expression indicating when insns that read
10221 memory in @var{mem_mode}, an integral mode narrower than a word, set the
10222 bits outside of @var{mem_mode} to be either the sign-extension or the
10223 zero-extension of the data read. Return @code{SIGN_EXTEND} for values
10224 of @var{mem_mode} for which the
10225 insn sign-extends, @code{ZERO_EXTEND} for which it zero-extends, and
10226 @code{UNKNOWN} for other modes.
10228 This macro is not called with @var{mem_mode} non-integral or with a width
10229 greater than or equal to @code{BITS_PER_WORD}, so you may return any
10230 value in this case. Do not define this macro if it would always return
10231 @code{UNKNOWN}. On machines where this macro is defined, you will normally
10232 define it as the constant @code{SIGN_EXTEND} or @code{ZERO_EXTEND}.
10234 You may return a non-@code{UNKNOWN} value even if for some hard registers
10235 the sign extension is not performed, if for the @code{REGNO_REG_CLASS}
10236 of these hard registers @code{CANNOT_CHANGE_MODE_CLASS} returns nonzero
10237 when the @var{from} mode is @var{mem_mode} and the @var{to} mode is any
10238 integral mode larger than this but not larger than @code{word_mode}.
10240 You must return @code{UNKNOWN} if for some hard registers that allow this
10241 mode, @code{CANNOT_CHANGE_MODE_CLASS} says that they cannot change to
10242 @code{word_mode}, but that they can change to another integral mode that
10243 is larger then @var{mem_mode} but still smaller than @code{word_mode}.
10246 @defmac SHORT_IMMEDIATES_SIGN_EXTEND
10247 Define this macro if loading short immediate values into registers sign
10251 @defmac FIXUNS_TRUNC_LIKE_FIX_TRUNC
10252 Define this macro if the same instructions that convert a floating
10253 point number to a signed fixed point number also convert validly to an
10257 @hook TARGET_MIN_DIVISIONS_FOR_RECIP_MUL
10258 When @option{-ffast-math} is in effect, GCC tries to optimize
10259 divisions by the same divisor, by turning them into multiplications by
10260 the reciprocal. This target hook specifies the minimum number of divisions
10261 that should be there for GCC to perform the optimization for a variable
10262 of mode @var{mode}. The default implementation returns 3 if the machine
10263 has an instruction for the division, and 2 if it does not.
10267 The maximum number of bytes that a single instruction can move quickly
10268 between memory and registers or between two memory locations.
10271 @defmac MAX_MOVE_MAX
10272 The maximum number of bytes that a single instruction can move quickly
10273 between memory and registers or between two memory locations. If this
10274 is undefined, the default is @code{MOVE_MAX}. Otherwise, it is the
10275 constant value that is the largest value that @code{MOVE_MAX} can have
10279 @defmac SHIFT_COUNT_TRUNCATED
10280 A C expression that is nonzero if on this machine the number of bits
10281 actually used for the count of a shift operation is equal to the number
10282 of bits needed to represent the size of the object being shifted. When
10283 this macro is nonzero, the compiler will assume that it is safe to omit
10284 a sign-extend, zero-extend, and certain bitwise `and' instructions that
10285 truncates the count of a shift operation. On machines that have
10286 instructions that act on bit-fields at variable positions, which may
10287 include `bit test' instructions, a nonzero @code{SHIFT_COUNT_TRUNCATED}
10288 also enables deletion of truncations of the values that serve as
10289 arguments to bit-field instructions.
10291 If both types of instructions truncate the count (for shifts) and
10292 position (for bit-field operations), or if no variable-position bit-field
10293 instructions exist, you should define this macro.
10295 However, on some machines, such as the 80386 and the 680x0, truncation
10296 only applies to shift operations and not the (real or pretended)
10297 bit-field operations. Define @code{SHIFT_COUNT_TRUNCATED} to be zero on
10298 such machines. Instead, add patterns to the @file{md} file that include
10299 the implied truncation of the shift instructions.
10301 You need not define this macro if it would always have the value of zero.
10304 @anchor{TARGET_SHIFT_TRUNCATION_MASK}
10305 @hook TARGET_SHIFT_TRUNCATION_MASK
10306 This function describes how the standard shift patterns for @var{mode}
10307 deal with shifts by negative amounts or by more than the width of the mode.
10308 @xref{shift patterns}.
10310 On many machines, the shift patterns will apply a mask @var{m} to the
10311 shift count, meaning that a fixed-width shift of @var{x} by @var{y} is
10312 equivalent to an arbitrary-width shift of @var{x} by @var{y & m}. If
10313 this is true for mode @var{mode}, the function should return @var{m},
10314 otherwise it should return 0. A return value of 0 indicates that no
10315 particular behavior is guaranteed.
10317 Note that, unlike @code{SHIFT_COUNT_TRUNCATED}, this function does
10318 @emph{not} apply to general shift rtxes; it applies only to instructions
10319 that are generated by the named shift patterns.
10321 The default implementation of this function returns
10322 @code{GET_MODE_BITSIZE (@var{mode}) - 1} if @code{SHIFT_COUNT_TRUNCATED}
10323 and 0 otherwise. This definition is always safe, but if
10324 @code{SHIFT_COUNT_TRUNCATED} is false, and some shift patterns
10325 nevertheless truncate the shift count, you may get better code
10329 @defmac TRULY_NOOP_TRUNCATION (@var{outprec}, @var{inprec})
10330 A C expression which is nonzero if on this machine it is safe to
10331 ``convert'' an integer of @var{inprec} bits to one of @var{outprec}
10332 bits (where @var{outprec} is smaller than @var{inprec}) by merely
10333 operating on it as if it had only @var{outprec} bits.
10335 On many machines, this expression can be 1.
10337 @c rearranged this, removed the phrase "it is reported that". this was
10338 @c to fix an overfull hbox. --mew 10feb93
10339 When @code{TRULY_NOOP_TRUNCATION} returns 1 for a pair of sizes for
10340 modes for which @code{MODES_TIEABLE_P} is 0, suboptimal code can result.
10341 If this is the case, making @code{TRULY_NOOP_TRUNCATION} return 0 in
10342 such cases may improve things.
10345 @hook TARGET_MODE_REP_EXTENDED
10346 The representation of an integral mode can be such that the values
10347 are always extended to a wider integral mode. Return
10348 @code{SIGN_EXTEND} if values of @var{mode} are represented in
10349 sign-extended form to @var{rep_mode}. Return @code{UNKNOWN}
10350 otherwise. (Currently, none of the targets use zero-extended
10351 representation this way so unlike @code{LOAD_EXTEND_OP},
10352 @code{TARGET_MODE_REP_EXTENDED} is expected to return either
10353 @code{SIGN_EXTEND} or @code{UNKNOWN}. Also no target extends
10354 @var{mode} to @var{rep_mode} so that @var{rep_mode} is not the next
10355 widest integral mode and currently we take advantage of this fact.)
10357 Similarly to @code{LOAD_EXTEND_OP} you may return a non-@code{UNKNOWN}
10358 value even if the extension is not performed on certain hard registers
10359 as long as for the @code{REGNO_REG_CLASS} of these hard registers
10360 @code{CANNOT_CHANGE_MODE_CLASS} returns nonzero.
10362 Note that @code{TARGET_MODE_REP_EXTENDED} and @code{LOAD_EXTEND_OP}
10363 describe two related properties. If you define
10364 @code{TARGET_MODE_REP_EXTENDED (mode, word_mode)} you probably also want
10365 to define @code{LOAD_EXTEND_OP (mode)} to return the same type of
10368 In order to enforce the representation of @code{mode},
10369 @code{TRULY_NOOP_TRUNCATION} should return false when truncating to
10373 @defmac STORE_FLAG_VALUE
10374 A C expression describing the value returned by a comparison operator
10375 with an integral mode and stored by a store-flag instruction
10376 (@samp{cstore@var{mode}4}) when the condition is true. This description must
10377 apply to @emph{all} the @samp{cstore@var{mode}4} patterns and all the
10378 comparison operators whose results have a @code{MODE_INT} mode.
10380 A value of 1 or @minus{}1 means that the instruction implementing the
10381 comparison operator returns exactly 1 or @minus{}1 when the comparison is true
10382 and 0 when the comparison is false. Otherwise, the value indicates
10383 which bits of the result are guaranteed to be 1 when the comparison is
10384 true. This value is interpreted in the mode of the comparison
10385 operation, which is given by the mode of the first operand in the
10386 @samp{cstore@var{mode}4} pattern. Either the low bit or the sign bit of
10387 @code{STORE_FLAG_VALUE} be on. Presently, only those bits are used by
10390 If @code{STORE_FLAG_VALUE} is neither 1 or @minus{}1, the compiler will
10391 generate code that depends only on the specified bits. It can also
10392 replace comparison operators with equivalent operations if they cause
10393 the required bits to be set, even if the remaining bits are undefined.
10394 For example, on a machine whose comparison operators return an
10395 @code{SImode} value and where @code{STORE_FLAG_VALUE} is defined as
10396 @samp{0x80000000}, saying that just the sign bit is relevant, the
10400 (ne:SI (and:SI @var{x} (const_int @var{power-of-2})) (const_int 0))
10404 can be converted to
10407 (ashift:SI @var{x} (const_int @var{n}))
10411 where @var{n} is the appropriate shift count to move the bit being
10412 tested into the sign bit.
10414 There is no way to describe a machine that always sets the low-order bit
10415 for a true value, but does not guarantee the value of any other bits,
10416 but we do not know of any machine that has such an instruction. If you
10417 are trying to port GCC to such a machine, include an instruction to
10418 perform a logical-and of the result with 1 in the pattern for the
10419 comparison operators and let us know at @email{gcc@@gcc.gnu.org}.
10421 Often, a machine will have multiple instructions that obtain a value
10422 from a comparison (or the condition codes). Here are rules to guide the
10423 choice of value for @code{STORE_FLAG_VALUE}, and hence the instructions
10428 Use the shortest sequence that yields a valid definition for
10429 @code{STORE_FLAG_VALUE}. It is more efficient for the compiler to
10430 ``normalize'' the value (convert it to, e.g., 1 or 0) than for the
10431 comparison operators to do so because there may be opportunities to
10432 combine the normalization with other operations.
10435 For equal-length sequences, use a value of 1 or @minus{}1, with @minus{}1 being
10436 slightly preferred on machines with expensive jumps and 1 preferred on
10440 As a second choice, choose a value of @samp{0x80000001} if instructions
10441 exist that set both the sign and low-order bits but do not define the
10445 Otherwise, use a value of @samp{0x80000000}.
10448 Many machines can produce both the value chosen for
10449 @code{STORE_FLAG_VALUE} and its negation in the same number of
10450 instructions. On those machines, you should also define a pattern for
10451 those cases, e.g., one matching
10454 (set @var{A} (neg:@var{m} (ne:@var{m} @var{B} @var{C})))
10457 Some machines can also perform @code{and} or @code{plus} operations on
10458 condition code values with less instructions than the corresponding
10459 @samp{cstore@var{mode}4} insn followed by @code{and} or @code{plus}. On those
10460 machines, define the appropriate patterns. Use the names @code{incscc}
10461 and @code{decscc}, respectively, for the patterns which perform
10462 @code{plus} or @code{minus} operations on condition code values. See
10463 @file{rs6000.md} for some examples. The GNU Superoptimizer can be used to
10464 find such instruction sequences on other machines.
10466 If this macro is not defined, the default value, 1, is used. You need
10467 not define @code{STORE_FLAG_VALUE} if the machine has no store-flag
10468 instructions, or if the value generated by these instructions is 1.
10471 @defmac FLOAT_STORE_FLAG_VALUE (@var{mode})
10472 A C expression that gives a nonzero @code{REAL_VALUE_TYPE} value that is
10473 returned when comparison operators with floating-point results are true.
10474 Define this macro on machines that have comparison operations that return
10475 floating-point values. If there are no such operations, do not define
10479 @defmac VECTOR_STORE_FLAG_VALUE (@var{mode})
10480 A C expression that gives a rtx representing the nonzero true element
10481 for vector comparisons. The returned rtx should be valid for the inner
10482 mode of @var{mode} which is guaranteed to be a vector mode. Define
10483 this macro on machines that have vector comparison operations that
10484 return a vector result. If there are no such operations, do not define
10485 this macro. Typically, this macro is defined as @code{const1_rtx} or
10486 @code{constm1_rtx}. This macro may return @code{NULL_RTX} to prevent
10487 the compiler optimizing such vector comparison operations for the
10491 @defmac CLZ_DEFINED_VALUE_AT_ZERO (@var{mode}, @var{value})
10492 @defmacx CTZ_DEFINED_VALUE_AT_ZERO (@var{mode}, @var{value})
10493 A C expression that indicates whether the architecture defines a value
10494 for @code{clz} or @code{ctz} with a zero operand.
10495 A result of @code{0} indicates the value is undefined.
10496 If the value is defined for only the RTL expression, the macro should
10497 evaluate to @code{1}; if the value applies also to the corresponding optab
10498 entry (which is normally the case if it expands directly into
10499 the corresponding RTL), then the macro should evaluate to @code{2}.
10500 In the cases where the value is defined, @var{value} should be set to
10503 If this macro is not defined, the value of @code{clz} or
10504 @code{ctz} at zero is assumed to be undefined.
10506 This macro must be defined if the target's expansion for @code{ffs}
10507 relies on a particular value to get correct results. Otherwise it
10508 is not necessary, though it may be used to optimize some corner cases, and
10509 to provide a default expansion for the @code{ffs} optab.
10511 Note that regardless of this macro the ``definedness'' of @code{clz}
10512 and @code{ctz} at zero do @emph{not} extend to the builtin functions
10513 visible to the user. Thus one may be free to adjust the value at will
10514 to match the target expansion of these operations without fear of
10519 An alias for the machine mode for pointers. On most machines, define
10520 this to be the integer mode corresponding to the width of a hardware
10521 pointer; @code{SImode} on 32-bit machine or @code{DImode} on 64-bit machines.
10522 On some machines you must define this to be one of the partial integer
10523 modes, such as @code{PSImode}.
10525 The width of @code{Pmode} must be at least as large as the value of
10526 @code{POINTER_SIZE}. If it is not equal, you must define the macro
10527 @code{POINTERS_EXTEND_UNSIGNED} to specify how pointers are extended
10531 @defmac FUNCTION_MODE
10532 An alias for the machine mode used for memory references to functions
10533 being called, in @code{call} RTL expressions. On most CISC machines,
10534 where an instruction can begin at any byte address, this should be
10535 @code{QImode}. On most RISC machines, where all instructions have fixed
10536 size and alignment, this should be a mode with the same size and alignment
10537 as the machine instruction words - typically @code{SImode} or @code{HImode}.
10540 @defmac STDC_0_IN_SYSTEM_HEADERS
10541 In normal operation, the preprocessor expands @code{__STDC__} to the
10542 constant 1, to signify that GCC conforms to ISO Standard C@. On some
10543 hosts, like Solaris, the system compiler uses a different convention,
10544 where @code{__STDC__} is normally 0, but is 1 if the user specifies
10545 strict conformance to the C Standard.
10547 Defining @code{STDC_0_IN_SYSTEM_HEADERS} makes GNU CPP follows the host
10548 convention when processing system header files, but when processing user
10549 files @code{__STDC__} will always expand to 1.
10552 @defmac NO_IMPLICIT_EXTERN_C
10553 Define this macro if the system header files support C++ as well as C@.
10554 This macro inhibits the usual method of using system header files in
10555 C++, which is to pretend that the file's contents are enclosed in
10556 @samp{extern "C" @{@dots{}@}}.
10561 @defmac REGISTER_TARGET_PRAGMAS ()
10562 Define this macro if you want to implement any target-specific pragmas.
10563 If defined, it is a C expression which makes a series of calls to
10564 @code{c_register_pragma} or @code{c_register_pragma_with_expansion}
10565 for each pragma. The macro may also do any
10566 setup required for the pragmas.
10568 The primary reason to define this macro is to provide compatibility with
10569 other compilers for the same target. In general, we discourage
10570 definition of target-specific pragmas for GCC@.
10572 If the pragma can be implemented by attributes then you should consider
10573 defining the target hook @samp{TARGET_INSERT_ATTRIBUTES} as well.
10575 Preprocessor macros that appear on pragma lines are not expanded. All
10576 @samp{#pragma} directives that do not match any registered pragma are
10577 silently ignored, unless the user specifies @option{-Wunknown-pragmas}.
10580 @deftypefun void c_register_pragma (const char *@var{space}, const char *@var{name}, void (*@var{callback}) (struct cpp_reader *))
10581 @deftypefunx void c_register_pragma_with_expansion (const char *@var{space}, const char *@var{name}, void (*@var{callback}) (struct cpp_reader *))
10583 Each call to @code{c_register_pragma} or
10584 @code{c_register_pragma_with_expansion} establishes one pragma. The
10585 @var{callback} routine will be called when the preprocessor encounters a
10589 #pragma [@var{space}] @var{name} @dots{}
10592 @var{space} is the case-sensitive namespace of the pragma, or
10593 @code{NULL} to put the pragma in the global namespace. The callback
10594 routine receives @var{pfile} as its first argument, which can be passed
10595 on to cpplib's functions if necessary. You can lex tokens after the
10596 @var{name} by calling @code{pragma_lex}. Tokens that are not read by the
10597 callback will be silently ignored. The end of the line is indicated by
10598 a token of type @code{CPP_EOF}. Macro expansion occurs on the
10599 arguments of pragmas registered with
10600 @code{c_register_pragma_with_expansion} but not on the arguments of
10601 pragmas registered with @code{c_register_pragma}.
10603 Note that the use of @code{pragma_lex} is specific to the C and C++
10604 compilers. It will not work in the Java or Fortran compilers, or any
10605 other language compilers for that matter. Thus if @code{pragma_lex} is going
10606 to be called from target-specific code, it must only be done so when
10607 building the C and C++ compilers. This can be done by defining the
10608 variables @code{c_target_objs} and @code{cxx_target_objs} in the
10609 target entry in the @file{config.gcc} file. These variables should name
10610 the target-specific, language-specific object file which contains the
10611 code that uses @code{pragma_lex}. Note it will also be necessary to add a
10612 rule to the makefile fragment pointed to by @code{tmake_file} that shows
10613 how to build this object file.
10616 @defmac HANDLE_PRAGMA_PACK_WITH_EXPANSION
10617 Define this macro if macros should be expanded in the
10618 arguments of @samp{#pragma pack}.
10621 @hook TARGET_HANDLE_PRAGMA_EXTERN_PREFIX
10623 @defmac TARGET_DEFAULT_PACK_STRUCT
10624 If your target requires a structure packing default other than 0 (meaning
10625 the machine default), define this macro to the necessary value (in bytes).
10626 This must be a value that would also be valid to use with
10627 @samp{#pragma pack()} (that is, a small power of two).
10630 @defmac DOLLARS_IN_IDENTIFIERS
10631 Define this macro to control use of the character @samp{$} in
10632 identifier names for the C family of languages. 0 means @samp{$} is
10633 not allowed by default; 1 means it is allowed. 1 is the default;
10634 there is no need to define this macro in that case.
10637 @defmac NO_DOLLAR_IN_LABEL
10638 Define this macro if the assembler does not accept the character
10639 @samp{$} in label names. By default constructors and destructors in
10640 G++ have @samp{$} in the identifiers. If this macro is defined,
10641 @samp{.} is used instead.
10644 @defmac NO_DOT_IN_LABEL
10645 Define this macro if the assembler does not accept the character
10646 @samp{.} in label names. By default constructors and destructors in G++
10647 have names that use @samp{.}. If this macro is defined, these names
10648 are rewritten to avoid @samp{.}.
10651 @defmac INSN_SETS_ARE_DELAYED (@var{insn})
10652 Define this macro as a C expression that is nonzero if it is safe for the
10653 delay slot scheduler to place instructions in the delay slot of @var{insn},
10654 even if they appear to use a resource set or clobbered in @var{insn}.
10655 @var{insn} is always a @code{jump_insn} or an @code{insn}; GCC knows that
10656 every @code{call_insn} has this behavior. On machines where some @code{insn}
10657 or @code{jump_insn} is really a function call and hence has this behavior,
10658 you should define this macro.
10660 You need not define this macro if it would always return zero.
10663 @defmac INSN_REFERENCES_ARE_DELAYED (@var{insn})
10664 Define this macro as a C expression that is nonzero if it is safe for the
10665 delay slot scheduler to place instructions in the delay slot of @var{insn},
10666 even if they appear to set or clobber a resource referenced in @var{insn}.
10667 @var{insn} is always a @code{jump_insn} or an @code{insn}. On machines where
10668 some @code{insn} or @code{jump_insn} is really a function call and its operands
10669 are registers whose use is actually in the subroutine it calls, you should
10670 define this macro. Doing so allows the delay slot scheduler to move
10671 instructions which copy arguments into the argument registers into the delay
10672 slot of @var{insn}.
10674 You need not define this macro if it would always return zero.
10677 @defmac MULTIPLE_SYMBOL_SPACES
10678 Define this macro as a C expression that is nonzero if, in some cases,
10679 global symbols from one translation unit may not be bound to undefined
10680 symbols in another translation unit without user intervention. For
10681 instance, under Microsoft Windows symbols must be explicitly imported
10682 from shared libraries (DLLs).
10684 You need not define this macro if it would always evaluate to zero.
10687 @hook TARGET_MD_ASM_CLOBBERS
10688 This target hook should add to @var{clobbers} @code{STRING_CST} trees for
10689 any hard regs the port wishes to automatically clobber for an asm.
10690 It should return the result of the last @code{tree_cons} used to add a
10691 clobber. The @var{outputs}, @var{inputs} and @var{clobber} lists are the
10692 corresponding parameters to the asm and may be inspected to avoid
10693 clobbering a register that is an input or output of the asm. You can use
10694 @code{tree_overlaps_hard_reg_set}, declared in @file{tree.h}, to test
10695 for overlap with regards to asm-declared registers.
10698 @defmac MATH_LIBRARY
10699 Define this macro as a C string constant for the linker argument to link
10700 in the system math library, minus the initial @samp{"-l"}, or
10701 @samp{""} if the target does not have a
10702 separate math library.
10704 You need only define this macro if the default of @samp{"m"} is wrong.
10707 @defmac LIBRARY_PATH_ENV
10708 Define this macro as a C string constant for the environment variable that
10709 specifies where the linker should look for libraries.
10711 You need only define this macro if the default of @samp{"LIBRARY_PATH"}
10715 @defmac TARGET_POSIX_IO
10716 Define this macro if the target supports the following POSIX@ file
10717 functions, access, mkdir and file locking with fcntl / F_SETLKW@.
10718 Defining @code{TARGET_POSIX_IO} will enable the test coverage code
10719 to use file locking when exiting a program, which avoids race conditions
10720 if the program has forked. It will also create directories at run-time
10721 for cross-profiling.
10724 @defmac MAX_CONDITIONAL_EXECUTE
10726 A C expression for the maximum number of instructions to execute via
10727 conditional execution instructions instead of a branch. A value of
10728 @code{BRANCH_COST}+1 is the default if the machine does not use cc0, and
10729 1 if it does use cc0.
10732 @defmac IFCVT_MODIFY_TESTS (@var{ce_info}, @var{true_expr}, @var{false_expr})
10733 Used if the target needs to perform machine-dependent modifications on the
10734 conditionals used for turning basic blocks into conditionally executed code.
10735 @var{ce_info} points to a data structure, @code{struct ce_if_block}, which
10736 contains information about the currently processed blocks. @var{true_expr}
10737 and @var{false_expr} are the tests that are used for converting the
10738 then-block and the else-block, respectively. Set either @var{true_expr} or
10739 @var{false_expr} to a null pointer if the tests cannot be converted.
10742 @defmac IFCVT_MODIFY_MULTIPLE_TESTS (@var{ce_info}, @var{bb}, @var{true_expr}, @var{false_expr})
10743 Like @code{IFCVT_MODIFY_TESTS}, but used when converting more complicated
10744 if-statements into conditions combined by @code{and} and @code{or} operations.
10745 @var{bb} contains the basic block that contains the test that is currently
10746 being processed and about to be turned into a condition.
10749 @defmac IFCVT_MODIFY_INSN (@var{ce_info}, @var{pattern}, @var{insn})
10750 A C expression to modify the @var{PATTERN} of an @var{INSN} that is to
10751 be converted to conditional execution format. @var{ce_info} points to
10752 a data structure, @code{struct ce_if_block}, which contains information
10753 about the currently processed blocks.
10756 @defmac IFCVT_MODIFY_FINAL (@var{ce_info})
10757 A C expression to perform any final machine dependent modifications in
10758 converting code to conditional execution. The involved basic blocks
10759 can be found in the @code{struct ce_if_block} structure that is pointed
10760 to by @var{ce_info}.
10763 @defmac IFCVT_MODIFY_CANCEL (@var{ce_info})
10764 A C expression to cancel any machine dependent modifications in
10765 converting code to conditional execution. The involved basic blocks
10766 can be found in the @code{struct ce_if_block} structure that is pointed
10767 to by @var{ce_info}.
10770 @defmac IFCVT_INIT_EXTRA_FIELDS (@var{ce_info})
10771 A C expression to initialize any extra fields in a @code{struct ce_if_block}
10772 structure, which are defined by the @code{IFCVT_EXTRA_FIELDS} macro.
10775 @defmac IFCVT_EXTRA_FIELDS
10776 If defined, it should expand to a set of field declarations that will be
10777 added to the @code{struct ce_if_block} structure. These should be initialized
10778 by the @code{IFCVT_INIT_EXTRA_FIELDS} macro.
10781 @hook TARGET_MACHINE_DEPENDENT_REORG
10782 If non-null, this hook performs a target-specific pass over the
10783 instruction stream. The compiler will run it at all optimization levels,
10784 just before the point at which it normally does delayed-branch scheduling.
10786 The exact purpose of the hook varies from target to target. Some use
10787 it to do transformations that are necessary for correctness, such as
10788 laying out in-function constant pools or avoiding hardware hazards.
10789 Others use it as an opportunity to do some machine-dependent optimizations.
10791 You need not implement the hook if it has nothing to do. The default
10792 definition is null.
10795 @hook TARGET_INIT_BUILTINS
10796 Define this hook if you have any machine-specific built-in functions
10797 that need to be defined. It should be a function that performs the
10800 Machine specific built-in functions can be useful to expand special machine
10801 instructions that would otherwise not normally be generated because
10802 they have no equivalent in the source language (for example, SIMD vector
10803 instructions or prefetch instructions).
10805 To create a built-in function, call the function
10806 @code{lang_hooks.builtin_function}
10807 which is defined by the language front end. You can use any type nodes set
10808 up by @code{build_common_tree_nodes};
10809 only language front ends that use those two functions will call
10810 @samp{TARGET_INIT_BUILTINS}.
10813 @hook TARGET_BUILTIN_DECL
10814 Define this hook if you have any machine-specific built-in functions
10815 that need to be defined. It should be a function that returns the
10816 builtin function declaration for the builtin function code @var{code}.
10817 If there is no such builtin and it cannot be initialized at this time
10818 if @var{initialize_p} is true the function should return @code{NULL_TREE}.
10819 If @var{code} is out of range the function should return
10820 @code{error_mark_node}.
10823 @hook TARGET_EXPAND_BUILTIN
10825 Expand a call to a machine specific built-in function that was set up by
10826 @samp{TARGET_INIT_BUILTINS}. @var{exp} is the expression for the
10827 function call; the result should go to @var{target} if that is
10828 convenient, and have mode @var{mode} if that is convenient.
10829 @var{subtarget} may be used as the target for computing one of
10830 @var{exp}'s operands. @var{ignore} is nonzero if the value is to be
10831 ignored. This function should return the result of the call to the
10835 @hook TARGET_RESOLVE_OVERLOADED_BUILTIN
10836 Select a replacement for a machine specific built-in function that
10837 was set up by @samp{TARGET_INIT_BUILTINS}. This is done
10838 @emph{before} regular type checking, and so allows the target to
10839 implement a crude form of function overloading. @var{fndecl} is the
10840 declaration of the built-in function. @var{arglist} is the list of
10841 arguments passed to the built-in function. The result is a
10842 complete expression that implements the operation, usually
10843 another @code{CALL_EXPR}.
10844 @var{arglist} really has type @samp{VEC(tree,gc)*}
10847 @hook TARGET_FOLD_BUILTIN
10848 Fold a call to a machine specific built-in function that was set up by
10849 @samp{TARGET_INIT_BUILTINS}. @var{fndecl} is the declaration of the
10850 built-in function. @var{n_args} is the number of arguments passed to
10851 the function; the arguments themselves are pointed to by @var{argp}.
10852 The result is another tree containing a simplified expression for the
10853 call's result. If @var{ignore} is true the value will be ignored.
10856 @hook TARGET_INVALID_WITHIN_DOLOOP
10858 Take an instruction in @var{insn} and return NULL if it is valid within a
10859 low-overhead loop, otherwise return a string explaining why doloop
10860 could not be applied.
10862 Many targets use special registers for low-overhead looping. For any
10863 instruction that clobbers these this function should return a string indicating
10864 the reason why the doloop could not be applied.
10865 By default, the RTL loop optimizer does not use a present doloop pattern for
10866 loops containing function calls or branch on table instructions.
10869 @defmac MD_CAN_REDIRECT_BRANCH (@var{branch1}, @var{branch2})
10871 Take a branch insn in @var{branch1} and another in @var{branch2}.
10872 Return true if redirecting @var{branch1} to the destination of
10873 @var{branch2} is possible.
10875 On some targets, branches may have a limited range. Optimizing the
10876 filling of delay slots can result in branches being redirected, and this
10877 may in turn cause a branch offset to overflow.
10880 @hook TARGET_COMMUTATIVE_P
10881 This target hook returns @code{true} if @var{x} is considered to be commutative.
10882 Usually, this is just COMMUTATIVE_P (@var{x}), but the HP PA doesn't consider
10883 PLUS to be commutative inside a MEM@. @var{outer_code} is the rtx code
10884 of the enclosing rtl, if known, otherwise it is UNKNOWN.
10887 @hook TARGET_ALLOCATE_INITIAL_VALUE
10889 When the initial value of a hard register has been copied in a pseudo
10890 register, it is often not necessary to actually allocate another register
10891 to this pseudo register, because the original hard register or a stack slot
10892 it has been saved into can be used. @code{TARGET_ALLOCATE_INITIAL_VALUE}
10893 is called at the start of register allocation once for each hard register
10894 that had its initial value copied by using
10895 @code{get_func_hard_reg_initial_val} or @code{get_hard_reg_initial_val}.
10896 Possible values are @code{NULL_RTX}, if you don't want
10897 to do any special allocation, a @code{REG} rtx---that would typically be
10898 the hard register itself, if it is known not to be clobbered---or a
10900 If you are returning a @code{MEM}, this is only a hint for the allocator;
10901 it might decide to use another register anyways.
10902 You may use @code{current_function_leaf_function} in the hook, functions
10903 that use @code{REG_N_SETS}, to determine if the hard
10904 register in question will not be clobbered.
10905 The default value of this hook is @code{NULL}, which disables any special
10909 @hook TARGET_UNSPEC_MAY_TRAP_P
10910 This target hook returns nonzero if @var{x}, an @code{unspec} or
10911 @code{unspec_volatile} operation, might cause a trap. Targets can use
10912 this hook to enhance precision of analysis for @code{unspec} and
10913 @code{unspec_volatile} operations. You may call @code{may_trap_p_1}
10914 to analyze inner elements of @var{x} in which case @var{flags} should be
10918 @hook TARGET_SET_CURRENT_FUNCTION
10919 The compiler invokes this hook whenever it changes its current function
10920 context (@code{cfun}). You can define this function if
10921 the back end needs to perform any initialization or reset actions on a
10922 per-function basis. For example, it may be used to implement function
10923 attributes that affect register usage or code generation patterns.
10924 The argument @var{decl} is the declaration for the new function context,
10925 and may be null to indicate that the compiler has left a function context
10926 and is returning to processing at the top level.
10927 The default hook function does nothing.
10929 GCC sets @code{cfun} to a dummy function context during initialization of
10930 some parts of the back end. The hook function is not invoked in this
10931 situation; you need not worry about the hook being invoked recursively,
10932 or when the back end is in a partially-initialized state.
10933 @code{cfun} might be @code{NULL} to indicate processing at top level,
10934 outside of any function scope.
10937 @defmac TARGET_OBJECT_SUFFIX
10938 Define this macro to be a C string representing the suffix for object
10939 files on your target machine. If you do not define this macro, GCC will
10940 use @samp{.o} as the suffix for object files.
10943 @defmac TARGET_EXECUTABLE_SUFFIX
10944 Define this macro to be a C string representing the suffix to be
10945 automatically added to executable files on your target machine. If you
10946 do not define this macro, GCC will use the null string as the suffix for
10950 @defmac COLLECT_EXPORT_LIST
10951 If defined, @code{collect2} will scan the individual object files
10952 specified on its command line and create an export list for the linker.
10953 Define this macro for systems like AIX, where the linker discards
10954 object files that are not referenced from @code{main} and uses export
10958 @defmac MODIFY_JNI_METHOD_CALL (@var{mdecl})
10959 Define this macro to a C expression representing a variant of the
10960 method call @var{mdecl}, if Java Native Interface (JNI) methods
10961 must be invoked differently from other methods on your target.
10962 For example, on 32-bit Microsoft Windows, JNI methods must be invoked using
10963 the @code{stdcall} calling convention and this macro is then
10964 defined as this expression:
10967 build_type_attribute_variant (@var{mdecl},
10969 (get_identifier ("stdcall"),
10974 @hook TARGET_CANNOT_MODIFY_JUMPS_P
10975 This target hook returns @code{true} past the point in which new jump
10976 instructions could be created. On machines that require a register for
10977 every jump such as the SHmedia ISA of SH5, this point would typically be
10978 reload, so this target hook should be defined to a function such as:
10982 cannot_modify_jumps_past_reload_p ()
10984 return (reload_completed || reload_in_progress);
10989 @hook TARGET_BRANCH_TARGET_REGISTER_CLASS
10990 This target hook returns a register class for which branch target register
10991 optimizations should be applied. All registers in this class should be
10992 usable interchangeably. After reload, registers in this class will be
10993 re-allocated and loads will be hoisted out of loops and be subjected
10994 to inter-block scheduling.
10997 @hook TARGET_BRANCH_TARGET_REGISTER_CALLEE_SAVED
10998 Branch target register optimization will by default exclude callee-saved
11000 that are not already live during the current function; if this target hook
11001 returns true, they will be included. The target code must than make sure
11002 that all target registers in the class returned by
11003 @samp{TARGET_BRANCH_TARGET_REGISTER_CLASS} that might need saving are
11004 saved. @var{after_prologue_epilogue_gen} indicates if prologues and
11005 epilogues have already been generated. Note, even if you only return
11006 true when @var{after_prologue_epilogue_gen} is false, you still are likely
11007 to have to make special provisions in @code{INITIAL_ELIMINATION_OFFSET}
11008 to reserve space for caller-saved target registers.
11011 @hook TARGET_HAVE_CONDITIONAL_EXECUTION
11012 This target hook returns true if the target supports conditional execution.
11013 This target hook is required only when the target has several different
11014 modes and they have different conditional execution capability, such as ARM.
11017 @hook TARGET_LOOP_UNROLL_ADJUST
11018 This target hook returns a new value for the number of times @var{loop}
11019 should be unrolled. The parameter @var{nunroll} is the number of times
11020 the loop is to be unrolled. The parameter @var{loop} is a pointer to
11021 the loop, which is going to be checked for unrolling. This target hook
11022 is required only when the target has special constraints like maximum
11023 number of memory accesses.
11026 @defmac POWI_MAX_MULTS
11027 If defined, this macro is interpreted as a signed integer C expression
11028 that specifies the maximum number of floating point multiplications
11029 that should be emitted when expanding exponentiation by an integer
11030 constant inline. When this value is defined, exponentiation requiring
11031 more than this number of multiplications is implemented by calling the
11032 system library's @code{pow}, @code{powf} or @code{powl} routines.
11033 The default value places no upper bound on the multiplication count.
11036 @deftypefn Macro void TARGET_EXTRA_INCLUDES (const char *@var{sysroot}, const char *@var{iprefix}, int @var{stdinc})
11037 This target hook should register any extra include files for the
11038 target. The parameter @var{stdinc} indicates if normal include files
11039 are present. The parameter @var{sysroot} is the system root directory.
11040 The parameter @var{iprefix} is the prefix for the gcc directory.
11043 @deftypefn Macro void TARGET_EXTRA_PRE_INCLUDES (const char *@var{sysroot}, const char *@var{iprefix}, int @var{stdinc})
11044 This target hook should register any extra include files for the
11045 target before any standard headers. The parameter @var{stdinc}
11046 indicates if normal include files are present. The parameter
11047 @var{sysroot} is the system root directory. The parameter
11048 @var{iprefix} is the prefix for the gcc directory.
11051 @deftypefn Macro void TARGET_OPTF (char *@var{path})
11052 This target hook should register special include paths for the target.
11053 The parameter @var{path} is the include to register. On Darwin
11054 systems, this is used for Framework includes, which have semantics
11055 that are different from @option{-I}.
11058 @defmac bool TARGET_USE_LOCAL_THUNK_ALIAS_P (tree @var{fndecl})
11059 This target macro returns @code{true} if it is safe to use a local alias
11060 for a virtual function @var{fndecl} when constructing thunks,
11061 @code{false} otherwise. By default, the macro returns @code{true} for all
11062 functions, if a target supports aliases (i.e.@: defines
11063 @code{ASM_OUTPUT_DEF}), @code{false} otherwise,
11066 @defmac TARGET_FORMAT_TYPES
11067 If defined, this macro is the name of a global variable containing
11068 target-specific format checking information for the @option{-Wformat}
11069 option. The default is to have no target-specific format checks.
11072 @defmac TARGET_N_FORMAT_TYPES
11073 If defined, this macro is the number of entries in
11074 @code{TARGET_FORMAT_TYPES}.
11077 @defmac TARGET_OVERRIDES_FORMAT_ATTRIBUTES
11078 If defined, this macro is the name of a global variable containing
11079 target-specific format overrides for the @option{-Wformat} option. The
11080 default is to have no target-specific format overrides. If defined,
11081 @code{TARGET_FORMAT_TYPES} must be defined, too.
11084 @defmac TARGET_OVERRIDES_FORMAT_ATTRIBUTES_COUNT
11085 If defined, this macro specifies the number of entries in
11086 @code{TARGET_OVERRIDES_FORMAT_ATTRIBUTES}.
11089 @defmac TARGET_OVERRIDES_FORMAT_INIT
11090 If defined, this macro specifies the optional initialization
11091 routine for target specific customizations of the system printf
11092 and scanf formatter settings.
11095 @hook TARGET_RELAXED_ORDERING
11096 If set to @code{true}, means that the target's memory model does not
11097 guarantee that loads which do not depend on one another will access
11098 main memory in the order of the instruction stream; if ordering is
11099 important, an explicit memory barrier must be used. This is true of
11100 many recent processors which implement a policy of ``relaxed,''
11101 ``weak,'' or ``release'' memory consistency, such as Alpha, PowerPC,
11102 and ia64. The default is @code{false}.
11105 @hook TARGET_INVALID_ARG_FOR_UNPROTOTYPED_FN
11106 If defined, this macro returns the diagnostic message when it is
11107 illegal to pass argument @var{val} to function @var{funcdecl}
11108 with prototype @var{typelist}.
11111 @hook TARGET_INVALID_CONVERSION
11112 If defined, this macro returns the diagnostic message when it is
11113 invalid to convert from @var{fromtype} to @var{totype}, or @code{NULL}
11114 if validity should be determined by the front end.
11117 @hook TARGET_INVALID_UNARY_OP
11118 If defined, this macro returns the diagnostic message when it is
11119 invalid to apply operation @var{op} (where unary plus is denoted by
11120 @code{CONVERT_EXPR}) to an operand of type @var{type}, or @code{NULL}
11121 if validity should be determined by the front end.
11124 @hook TARGET_INVALID_BINARY_OP
11125 If defined, this macro returns the diagnostic message when it is
11126 invalid to apply operation @var{op} to operands of types @var{type1}
11127 and @var{type2}, or @code{NULL} if validity should be determined by
11131 @hook TARGET_INVALID_PARAMETER_TYPE
11132 If defined, this macro returns the diagnostic message when it is
11133 invalid for functions to include parameters of type @var{type},
11134 or @code{NULL} if validity should be determined by
11135 the front end. This is currently used only by the C and C++ front ends.
11138 @hook TARGET_INVALID_RETURN_TYPE
11139 If defined, this macro returns the diagnostic message when it is
11140 invalid for functions to have return type @var{type},
11141 or @code{NULL} if validity should be determined by
11142 the front end. This is currently used only by the C and C++ front ends.
11145 @hook TARGET_PROMOTED_TYPE
11146 If defined, this target hook returns the type to which values of
11147 @var{type} should be promoted when they appear in expressions,
11148 analogous to the integer promotions, or @code{NULL_TREE} to use the
11149 front end's normal promotion rules. This hook is useful when there are
11150 target-specific types with special promotion rules.
11151 This is currently used only by the C and C++ front ends.
11154 @hook TARGET_CONVERT_TO_TYPE
11155 If defined, this hook returns the result of converting @var{expr} to
11156 @var{type}. It should return the converted expression,
11157 or @code{NULL_TREE} to apply the front end's normal conversion rules.
11158 This hook is useful when there are target-specific types with special
11160 This is currently used only by the C and C++ front ends.
11163 @defmac TARGET_USE_JCR_SECTION
11164 This macro determines whether to use the JCR section to register Java
11165 classes. By default, TARGET_USE_JCR_SECTION is defined to 1 if both
11166 SUPPORTS_WEAK and TARGET_HAVE_NAMED_SECTIONS are true, else 0.
11170 This macro determines the size of the objective C jump buffer for the
11171 NeXT runtime. By default, OBJC_JBLEN is defined to an innocuous value.
11174 @defmac LIBGCC2_UNWIND_ATTRIBUTE
11175 Define this macro if any target-specific attributes need to be attached
11176 to the functions in @file{libgcc} that provide low-level support for
11177 call stack unwinding. It is used in declarations in @file{unwind-generic.h}
11178 and the associated definitions of those functions.
11181 @hook TARGET_UPDATE_STACK_BOUNDARY
11182 Define this macro to update the current function stack boundary if
11186 @hook TARGET_GET_DRAP_RTX
11187 This hook should return an rtx for Dynamic Realign Argument Pointer (DRAP) if a
11188 different argument pointer register is needed to access the function's
11189 argument list due to stack realignment. Return @code{NULL} if no DRAP
11193 @hook TARGET_ALLOCATE_STACK_SLOTS_FOR_ARGS
11194 When optimization is disabled, this hook indicates whether or not
11195 arguments should be allocated to stack slots. Normally, GCC allocates
11196 stacks slots for arguments when not optimizing in order to make
11197 debugging easier. However, when a function is declared with
11198 @code{__attribute__((naked))}, there is no stack frame, and the compiler
11199 cannot safely move arguments from the registers in which they are passed
11200 to the stack. Therefore, this hook should return true in general, but
11201 false for naked functions. The default implementation always returns true.
11204 @hook TARGET_CONST_ANCHOR
11205 On some architectures it can take multiple instructions to synthesize
11206 a constant. If there is another constant already in a register that
11207 is close enough in value then it is preferable that the new constant
11208 is computed from this register using immediate addition or
11209 subtraction. We accomplish this through CSE. Besides the value of
11210 the constant we also add a lower and an upper constant anchor to the
11211 available expressions. These are then queried when encountering new
11212 constants. The anchors are computed by rounding the constant up and
11213 down to a multiple of the value of @code{TARGET_CONST_ANCHOR}.
11214 @code{TARGET_CONST_ANCHOR} should be the maximum positive value
11215 accepted by immediate-add plus one. We currently assume that the
11216 value of @code{TARGET_CONST_ANCHOR} is a power of 2. For example, on
11217 MIPS, where add-immediate takes a 16-bit signed value,
11218 @code{TARGET_CONST_ANCHOR} is set to @samp{0x8000}. The default value
11219 is zero, which disables this optimization. @end deftypevr