1 /* Xstormy16 cpu description.
2 Copyright (C) 1997, 1998, 1999, 2000, 2001, 2002
3 Free Software Foundation, Inc.
4 Contributed by Red Hat, Inc.
6 This file is part of GNU CC.
8 GNU CC is free software; you can redistribute it and/or modify
9 it under the terms of the GNU General Public License as published by
10 the Free Software Foundation; either version 2, or (at your option)
13 GNU CC is distributed in the hope that it will be useful,
14 but WITHOUT ANY WARRANTY; without even the implied warranty of
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 GNU General Public License for more details.
18 You should have received a copy of the GNU General Public License
19 along with GNU CC; see the file COPYING. If not, write to
20 the Free Software Foundation, 59 Temple Place - Suite 330,
21 Boston, MA 02111-1307, USA. */
24 /* Driver configuration */
26 /* Defined in svr4.h. */
27 /* #define SWITCH_TAKES_ARG(CHAR) */
29 /* Defined in svr4.h. */
30 /* #define WORD_SWITCH_TAKES_ARG(NAME) */
32 /* Defined in svr4.h. */
36 /* Defined in svr4.h. */
37 /* #define ASM_FINAL_SPEC "" */
39 /* Defined in svr4.h. */
40 /* #define LINK_SPEC "" */
43 - If -msim is specified, everything is built and linked as for the sim.
44 - If -T is specified, that linker script is used, and it should provide
45 appropriate libraries.
46 - If neither is specified, everything is built as for the sim, but no
47 I/O support is assumed.
51 #define LIB_SPEC "-( -lc %{msim:-lsim}%{!msim:%{!T*:-lnosys}} -)"
53 /* Defined in svr4.h. */
55 #define STARTFILE_SPEC "crt0.o%s crti.o%s crtbegin.o%s"
57 /* Defined in svr4.h. */
59 #define ENDFILE_SPEC "crtend.o%s crtn.o%s"
61 /* Defined in svr4.h for host compilers. */
62 /* #define MD_EXEC_PREFIX "" */
64 /* Defined in svr4.h for host compilers. */
65 /* #define MD_STARTFILE_PREFIX "" */
68 /* Run-time target specifications */
70 #define CPP_PREDEFINES "-Dxstormy16 -Amachine=xstormy16 -D__INT_MAX__=32767"
72 /* This declaration should be present. */
73 extern int target_flags;
75 #define TARGET_SWITCHES \
76 {{ "sim", 0, "Provide libraries for the simulator" }, \
79 #define TARGET_VERSION fprintf (stderr, " (xstormy16 cpu core)");
81 #define CAN_DEBUG_WITHOUT_FP
86 #define BITS_BIG_ENDIAN 1
88 #define BYTES_BIG_ENDIAN 0
90 #define WORDS_BIG_ENDIAN 0
92 #define UNITS_PER_WORD 2
94 #define PROMOTE_MODE(MODE,UNSIGNEDP,TYPE) \
96 if (GET_MODE_CLASS (MODE) == MODE_INT \
97 && GET_MODE_SIZE (MODE) < 2) \
101 #define PROMOTE_FUNCTION_ARGS 1
103 #define PROMOTE_FUNCTION_RETURN 1
105 #define PARM_BOUNDARY 16
107 #define STACK_BOUNDARY 16
109 #define FUNCTION_BOUNDARY 16
111 #define BIGGEST_ALIGNMENT 16
113 /* Defined in svr4.h. */
114 /* #define MAX_OFILE_ALIGNMENT */
116 #define DATA_ALIGNMENT(TYPE, ALIGN) \
117 (TREE_CODE (TYPE) == ARRAY_TYPE \
118 && TYPE_MODE (TREE_TYPE (TYPE)) == QImode \
119 && (ALIGN) < BITS_PER_WORD ? BITS_PER_WORD : (ALIGN))
121 #define CONSTANT_ALIGNMENT(EXP, ALIGN) \
122 (TREE_CODE (EXP) == STRING_CST \
123 && (ALIGN) < BITS_PER_WORD ? BITS_PER_WORD : (ALIGN))
125 #define STRICT_ALIGNMENT 1
127 /* Defined in svr4.h. */
128 #define PCC_BITFIELD_TYPE_MATTERS 1
130 #define TARGET_FLOAT_FORMAT IEEE_FLOAT_FORMAT
133 /* Layout of Source Language Data Types */
135 #define INT_TYPE_SIZE 16
137 #define SHORT_TYPE_SIZE 16
139 #define LONG_TYPE_SIZE 32
141 #define LONG_LONG_TYPE_SIZE 64
143 #define FLOAT_TYPE_SIZE 32
145 #define DOUBLE_TYPE_SIZE 64
147 #define LONG_DOUBLE_TYPE_SIZE 64
149 #define DEFAULT_SIGNED_CHAR 0
151 /* Defined in svr4.h. */
152 #define SIZE_TYPE "unsigned int"
154 /* Defined in svr4.h. */
155 #define PTRDIFF_TYPE "int"
157 /* Defined in svr4.h, to "long int". */
158 /* #define WCHAR_TYPE "long int" */
160 /* Defined in svr4.h. */
161 #undef WCHAR_TYPE_SIZE
162 #define WCHAR_TYPE_SIZE 32
164 /* Define this macro if the type of Objective C selectors should be `int'.
166 If this macro is not defined, then selectors should have the type `struct
168 /* #define OBJC_INT_SELECTORS */
171 /* Register Basics */
173 /* Number of hardware registers known to the compiler. They receive numbers 0
174 through `FIRST_PSEUDO_REGISTER-1'; thus, the first pseudo register's number
175 really is assigned the number `FIRST_PSEUDO_REGISTER'. */
176 #define FIRST_PSEUDO_REGISTER 19
178 /* An initializer that says which registers are used for fixed purposes all
179 throughout the compiled code and are therefore not available for general
180 allocation. These would include the stack pointer, the frame pointer
181 (except on machines where that can be used as a general register when no
182 frame pointer is needed), the program counter on machines where that is
183 considered one of the addressable registers, and any other numbered register
186 This information is expressed as a sequence of numbers, separated by commas
187 and surrounded by braces. The Nth number is 1 if register N is fixed, 0
190 The table initialized from this macro, and the table initialized by the
191 following one, may be overridden at run time either automatically, by the
192 actions of the macro `CONDITIONAL_REGISTER_USAGE', or by the user with the
193 command options `-ffixed-REG', `-fcall-used-REG' and `-fcall-saved-REG'. */
194 #define FIXED_REGISTERS \
195 { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 1 }
197 /* Like `FIXED_REGISTERS' but has 1 for each register that is clobbered (in
198 general) by function calls as well as for fixed registers. This macro
199 therefore identifies the registers that are not available for general
200 allocation of values that must live across function calls.
202 If a register has 0 in `CALL_USED_REGISTERS', the compiler automatically
203 saves it on function entry and restores it on function exit, if the register
204 is used within the function. */
205 #define CALL_USED_REGISTERS \
206 { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 1, 1, 1, 1, 1 }
208 /* Zero or more C statements that may conditionally modify two variables
209 `fixed_regs' and `call_used_regs' (both of type `char []') after they have
210 been initialized from the two preceding macros.
212 This is necessary in case the fixed or call-clobbered registers depend on
215 You need not define this macro if it has no work to do.
217 If the usage of an entire class of registers depends on the target flags,
218 you may indicate this to GCC by using this macro to modify `fixed_regs' and
219 `call_used_regs' to 1 for each of the registers in the classes which should
220 not be used by GCC. Also define the macro `REG_CLASS_FROM_LETTER' to return
221 `NO_REGS' if it is called with a letter for a class that shouldn't be used.
223 (However, if this class is not included in `GENERAL_REGS' and all of the
224 insn patterns whose constraints permit this class are controlled by target
225 switches, then GCC will automatically avoid using these registers when the
226 target switches are opposed to them.) */
227 /* #define CONDITIONAL_REGISTER_USAGE */
229 /* If this macro is defined and has a nonzero value, it means that `setjmp' and
230 related functions fail to save the registers, or that `longjmp' fails to
231 restore them. To compensate, the compiler avoids putting variables in
232 registers in functions that use `setjmp'. */
233 /* #define NON_SAVING_SETJMP */
235 /* Define this macro if the target machine has register windows. This C
236 expression returns the register number as seen by the called function
237 corresponding to the register number OUT as seen by the calling function.
238 Return OUT if register number OUT is not an outbound register. */
239 /* #define INCOMING_REGNO(OUT) */
241 /* Define this macro if the target machine has register windows. This C
242 expression returns the register number as seen by the calling function
243 corresponding to the register number IN as seen by the called function.
244 Return IN if register number IN is not an inbound register. */
245 /* #define OUTGOING_REGNO(IN) */
248 /* Order of allocation of registers */
250 /* If defined, an initializer for a vector of integers, containing the numbers
251 of hard registers in the order in which GNU CC should prefer to use them
252 (from most preferred to least).
254 If this macro is not defined, registers are used lowest numbered first (all
257 One use of this macro is on machines where the highest numbered registers
258 must always be saved and the save-multiple-registers instruction supports
259 only sequences of consecutive registers. On such machines, define
260 `REG_ALLOC_ORDER' to be an initializer that lists the highest numbered
261 allocatable register first. */
262 #define REG_ALLOC_ORDER { 7, 6, 5, 4, 3, 2, 1, 0, 9, 8, 10, 11, 12, 13, 14, 15, 16 }
264 /* A C statement (sans semicolon) to choose the order in which to allocate hard
265 registers for pseudo-registers local to a basic block.
267 Store the desired register order in the array `reg_alloc_order'. Element 0
268 should be the register to allocate first; element 1, the next register; and
271 The macro body should not assume anything about the contents of
272 `reg_alloc_order' before execution of the macro.
274 On most machines, it is not necessary to define this macro. */
275 /* #define ORDER_REGS_FOR_LOCAL_ALLOC */
278 /* How Values Fit in Registers */
280 /* A C expression for the number of consecutive hard registers, starting at
281 register number REGNO, required to hold a value of mode MODE.
283 On a machine where all registers are exactly one word, a suitable definition
286 #define HARD_REGNO_NREGS(REGNO, MODE) \
287 ((GET_MODE_SIZE (MODE) + UNITS_PER_WORD - 1) \
288 / UNITS_PER_WORD)) */
289 #define HARD_REGNO_NREGS(REGNO, MODE) \
290 ((GET_MODE_SIZE (MODE) + UNITS_PER_WORD - 1) / UNITS_PER_WORD)
292 /* A C expression that is nonzero if it is permissible to store a value of mode
293 MODE in hard register number REGNO (or in several registers starting with
294 that one). For a machine where all registers are equivalent, a suitable
297 #define HARD_REGNO_MODE_OK(REGNO, MODE) 1
299 It is not necessary for this macro to check for the numbers of fixed
300 registers, because the allocation mechanism considers them to be always
303 On some machines, double-precision values must be kept in even/odd register
304 pairs. The way to implement that is to define this macro to reject odd
305 register numbers for such modes.
307 The minimum requirement for a mode to be OK in a register is that the
308 `movMODE' instruction pattern support moves between the register and any
309 other hard register for which the mode is OK; and that moving a value into
310 the register and back out not alter it.
312 Since the same instruction used to move `SImode' will work for all narrower
313 integer modes, it is not necessary on any machine for `HARD_REGNO_MODE_OK'
314 to distinguish between these modes, provided you define patterns `movhi',
315 etc., to take advantage of this. This is useful because of the interaction
316 between `HARD_REGNO_MODE_OK' and `MODES_TIEABLE_P'; it is very desirable for
317 all integer modes to be tieable.
319 Many machines have special registers for floating point arithmetic. Often
320 people assume that floating point machine modes are allowed only in floating
321 point registers. This is not true. Any registers that can hold integers
322 can safely *hold* a floating point machine mode, whether or not floating
323 arithmetic can be done on it in those registers. Integer move instructions
324 can be used to move the values.
326 On some machines, though, the converse is true: fixed-point machine modes
327 may not go in floating registers. This is true if the floating registers
328 normalize any value stored in them, because storing a non-floating value
329 there would garble it. In this case, `HARD_REGNO_MODE_OK' should reject
330 fixed-point machine modes in floating registers. But if the floating
331 registers do not automatically normalize, if you can store any bit pattern
332 in one and retrieve it unchanged without a trap, then any machine mode may
333 go in a floating register, so you can define this macro to say so.
335 The primary significance of special floating registers is rather that they
336 are the registers acceptable in floating point arithmetic instructions.
337 However, this is of no concern to `HARD_REGNO_MODE_OK'. You handle it by
338 writing the proper constraints for those instructions.
340 On some machines, the floating registers are especially slow to access, so
341 that it is better to store a value in a stack frame than in such a register
342 if floating point arithmetic is not being done. As long as the floating
343 registers are not in class `GENERAL_REGS', they will not be used unless some
344 pattern's constraint asks for one. */
345 #define HARD_REGNO_MODE_OK(REGNO, MODE) ((REGNO) != 16 || (MODE) == BImode)
347 /* A C expression that is nonzero if it is desirable to choose register
348 allocation so as to avoid move instructions between a value of mode MODE1
349 and a value of mode MODE2.
351 If `HARD_REGNO_MODE_OK (R, MODE1)' and `HARD_REGNO_MODE_OK (R, MODE2)' are
352 ever different for any R, then `MODES_TIEABLE_P (MODE1, MODE2)' must be
354 #define MODES_TIEABLE_P(MODE1, MODE2) ((MODE1) != BImode && (MODE2) != BImode)
356 /* Define this macro if the compiler should avoid copies to/from CCmode
357 registers. You should only define this macro if support fo copying to/from
358 CCmode is incomplete. */
359 /* #define AVOID_CCMODE_COPIES */
362 /* Handling Leaf Functions */
364 /* A C initializer for a vector, indexed by hard register number, which
365 contains 1 for a register that is allowable in a candidate for leaf function
368 If leaf function treatment involves renumbering the registers, then the
369 registers marked here should be the ones before renumbering--those that GNU
370 CC would ordinarily allocate. The registers which will actually be used in
371 the assembler code, after renumbering, should not be marked with 1 in this
374 Define this macro only if the target machine offers a way to optimize the
375 treatment of leaf functions. */
376 /* #define LEAF_REGISTERS */
378 /* A C expression whose value is the register number to which REGNO should be
379 renumbered, when a function is treated as a leaf function.
381 If REGNO is a register number which should not appear in a leaf function
382 before renumbering, then the expression should yield -1, which will cause
383 the compiler to abort.
385 Define this macro only if the target machine offers a way to optimize the
386 treatment of leaf functions, and registers need to be renumbered to do this. */
387 /* #define LEAF_REG_REMAP(REGNO) */
390 /* Registers That Form a Stack. */
392 /* Define this if the machine has any stack-like registers. */
393 /* #define STACK_REGS */
395 /* The number of the first stack-like register. This one is the top
397 /* #define FIRST_STACK_REG */
399 /* The number of the last stack-like register. This one is the
400 bottom of the stack. */
401 /* #define LAST_STACK_REG */
404 /* Register Classes */
406 /* An enumeral type that must be defined with all the register class names as
407 enumeral values. `NO_REGS' must be first. `ALL_REGS' must be the last
408 register class, followed by one more enumeral value, `LIM_REG_CLASSES',
409 which is not a register class but rather tells how many classes there are.
411 Each register class has a number, which is the value of casting the class
412 name to type `int'. The number serves as an index in many of the tables
430 /* The number of distinct register classes, defined as follows:
432 #define N_REG_CLASSES (int) LIM_REG_CLASSES */
433 #define N_REG_CLASSES ((int) LIM_REG_CLASSES)
435 /* An initializer containing the names of the register classes as C string
436 constants. These names are used in writing some of the debugging dumps. */
437 #define REG_CLASS_NAMES \
452 /* An initializer containing the contents of the register classes, as integers
453 which are bit masks. The Nth integer specifies the contents of class N.
454 The way the integer MASK is interpreted is that register R is in the class
455 if `MASK & (1 << R)' is 1.
457 When the machine has more than 32 registers, an integer does not suffice.
458 Then the integers are replaced by sub-initializers, braced groupings
459 containing several integers. Each sub-initializer must be suitable as an
460 initializer for the type `HARD_REG_SET' which is defined in
462 #define REG_CLASS_CONTENTS \
474 { (1 << FIRST_PSEUDO_REGISTER) - 1 } \
477 /* A C expression whose value is a register class containing hard register
478 REGNO. In general there is more than one such class; choose a class which
479 is "minimal", meaning that no smaller class also contains the register. */
480 #define REGNO_REG_CLASS(REGNO) \
481 ((REGNO) == 0 ? R0_REGS \
482 : (REGNO) == 1 ? R1_REGS \
483 : (REGNO) == 2 ? R2_REGS \
484 : (REGNO) < 8 ? EIGHT_REGS \
485 : (REGNO) == 8 ? R8_REGS \
486 : (REGNO) == 16 ? CARRY_REGS \
487 : (REGNO) <= 18 ? GENERAL_REGS \
490 /* A macro whose definition is the name of the class to which a valid base
491 register must belong. A base register is one used in an address which is
492 the register value plus a displacement. */
493 #define BASE_REG_CLASS GENERAL_REGS
495 /* A macro whose definition is the name of the class to which a valid index
496 register must belong. An index register is one used in an address where its
497 value is either multiplied by a scale factor or added to another register
498 (as well as added to a displacement). */
499 #define INDEX_REG_CLASS GENERAL_REGS
501 /* A C expression which defines the machine-dependent operand constraint
502 letters for register classes. If CHAR is such a letter, the value should be
503 the register class corresponding to it. Otherwise, the value should be
504 `NO_REGS'. The register letter `r', corresponding to class `GENERAL_REGS',
505 will not be passed to this macro; you do not need to handle it.
507 The following letters are unavailable, due to being used as
512 'I', 'J', 'K', 'L', 'M', 'N', 'O', 'P'
513 'Q', 'R', 'S', 'T', 'U'
515 'g', 'i', 'm', 'n', 'o', 'p', 'r', 's' */
517 #define REG_CLASS_FROM_LETTER(CHAR) \
518 ( (CHAR) == 'a' ? R0_REGS \
519 : (CHAR) == 'b' ? R1_REGS \
520 : (CHAR) == 'c' ? R2_REGS \
521 : (CHAR) == 'd' ? R8_REGS \
522 : (CHAR) == 'e' ? EIGHT_REGS \
523 : (CHAR) == 't' ? TWO_REGS \
524 : (CHAR) == 'y' ? CARRY_REGS \
525 : (CHAR) == 'z' ? ICALL_REGS \
528 /* A C expression which is nonzero if register number NUM is suitable for use
529 as a base register in operand addresses. It may be either a suitable hard
530 register or a pseudo register that has been allocated such a hard register. */
531 #define REGNO_OK_FOR_BASE_P(NUM) 1
533 /* A C expression which is nonzero if register number NUM is suitable for use
534 as an index register in operand addresses. It may be either a suitable hard
535 register or a pseudo register that has been allocated such a hard register.
537 The difference between an index register and a base register is that the
538 index register may be scaled. If an address involves the sum of two
539 registers, neither one of them scaled, then either one may be labeled the
540 "base" and the other the "index"; but whichever labeling is used must fit
541 the machine's constraints of which registers may serve in each capacity.
542 The compiler will try both labelings, looking for one that is valid, and
543 will reload one or both registers only if neither labeling works. */
544 #define REGNO_OK_FOR_INDEX_P(NUM) REGNO_OK_FOR_BASE_P (NUM)
546 /* A C expression that places additional restrictions on the register class to
547 use when it is necessary to copy value X into a register in class CLASS.
548 The value is a register class; perhaps CLASS, or perhaps another, smaller
549 class. On many machines, the following definition is safe:
551 #define PREFERRED_RELOAD_CLASS(X,CLASS) CLASS
553 Sometimes returning a more restrictive class makes better code. For
554 example, on the 68000, when X is an integer constant that is in range for a
555 `moveq' instruction, the value of this macro is always `DATA_REGS' as long
556 as CLASS includes the data registers. Requiring a data register guarantees
557 that a `moveq' will be used.
559 If X is a `const_double', by returning `NO_REGS' you can force X into a
560 memory constant. This is useful on certain machines where immediate
561 floating values cannot be loaded into certain kinds of registers.
563 This declaration must be present. */
564 #define PREFERRED_RELOAD_CLASS(X, CLASS) \
565 xstormy16_preferred_reload_class (X, CLASS)
567 /* Like `PREFERRED_RELOAD_CLASS', but for output reloads instead of input
568 reloads. If you don't define this macro, the default is to use CLASS,
570 #define PREFERRED_OUTPUT_RELOAD_CLASS(X, CLASS) \
571 xstormy16_preferred_reload_class (X, CLASS)
573 /* A C expression that places additional restrictions on the register class to
574 use when it is necessary to be able to hold a value of mode MODE in a reload
575 register for which class CLASS would ordinarily be used.
577 Unlike `PREFERRED_RELOAD_CLASS', this macro should be used when there are
578 certain modes that simply can't go in certain reload classes.
580 The value is a register class; perhaps CLASS, or perhaps another, smaller
583 Don't define this macro unless the target machine has limitations which
584 require the macro to do something nontrivial. */
585 /* #define LIMIT_RELOAD_CLASS(MODE, CLASS) */
587 /* Many machines have some registers that cannot be copied directly to or from
588 memory or even from other types of registers. An example is the `MQ'
589 register, which on most machines, can only be copied to or from general
590 registers, but not memory. Some machines allow copying all registers to and
591 from memory, but require a scratch register for stores to some memory
592 locations (e.g., those with symbolic address on the RT, and those with
593 certain symbolic address on the Sparc when compiling PIC). In some cases,
594 both an intermediate and a scratch register are required.
596 You should define these macros to indicate to the reload phase that it may
597 need to allocate at least one register for a reload in addition to the
598 register to contain the data. Specifically, if copying X to a register
599 CLASS in MODE requires an intermediate register, you should define
600 `SECONDARY_INPUT_RELOAD_CLASS' to return the largest register class all of
601 whose registers can be used as intermediate registers or scratch registers.
603 If copying a register CLASS in MODE to X requires an intermediate or scratch
604 register, `SECONDARY_OUTPUT_RELOAD_CLASS' should be defined to return the
605 largest register class required. If the requirements for input and output
606 reloads are the same, the macro `SECONDARY_RELOAD_CLASS' should be used
607 instead of defining both macros identically.
609 The values returned by these macros are often `GENERAL_REGS'. Return
610 `NO_REGS' if no spare register is needed; i.e., if X can be directly copied
611 to or from a register of CLASS in MODE without requiring a scratch register.
612 Do not define this macro if it would always return `NO_REGS'.
614 If a scratch register is required (either with or without an intermediate
615 register), you should define patterns for `reload_inM' or `reload_outM', as
616 required.. These patterns, which will normally be implemented with a
617 `define_expand', should be similar to the `movM' patterns, except that
618 operand 2 is the scratch register.
620 Define constraints for the reload register and scratch register that contain
621 a single register class. If the original reload register (whose class is
622 CLASS) can meet the constraint given in the pattern, the value returned by
623 these macros is used for the class of the scratch register. Otherwise, two
624 additional reload registers are required. Their classes are obtained from
625 the constraints in the insn pattern.
627 X might be a pseudo-register or a `subreg' of a pseudo-register, which could
628 either be in a hard register or in memory. Use `true_regnum' to find out;
629 it will return -1 if the pseudo is in memory and the hard register number if
632 These macros should not be used in the case where a particular class of
633 registers can only be copied to memory and not to another class of
634 registers. In that case, secondary reload registers are not needed and
635 would not be helpful. Instead, a stack location must be used to perform the
636 copy and the `movM' pattern should use memory as an intermediate storage.
637 This case often occurs between floating-point and general registers. */
639 /* This chip has the interesting property that only the first eight
640 registers can be moved to/from memory. */
641 #define SECONDARY_RELOAD_CLASS(CLASS, MODE, X) \
642 xstormy16_secondary_reload_class (CLASS, MODE, X)
644 /* #define SECONDARY_INPUT_RELOAD_CLASS(CLASS, MODE, X) */
645 /* #define SECONDARY_OUTPUT_RELOAD_CLASS(CLASS, MODE, X) */
647 /* Certain machines have the property that some registers cannot be copied to
648 some other registers without using memory. Define this macro on those
649 machines to be a C expression that is non-zero if objects of mode M in
650 registers of CLASS1 can only be copied to registers of class CLASS2 by
651 storing a register of CLASS1 into memory and loading that memory location
652 into a register of CLASS2.
654 Do not define this macro if its value would always be zero. */
655 /* #define SECONDARY_MEMORY_NEEDED(CLASS1, CLASS2, M) */
657 /* Normally when `SECONDARY_MEMORY_NEEDED' is defined, the compiler allocates a
658 stack slot for a memory location needed for register copies. If this macro
659 is defined, the compiler instead uses the memory location defined by this
662 Do not define this macro if you do not define
663 `SECONDARY_MEMORY_NEEDED'. */
664 /* #define SECONDARY_MEMORY_NEEDED_RTX(MODE) */
666 /* When the compiler needs a secondary memory location to copy between two
667 registers of mode MODE, it normally allocates sufficient memory to hold a
668 quantity of `BITS_PER_WORD' bits and performs the store and load operations
669 in a mode that many bits wide and whose class is the same as that of MODE.
671 This is right thing to do on most machines because it ensures that all bits
672 of the register are copied and prevents accesses to the registers in a
673 narrower mode, which some machines prohibit for floating-point registers.
675 However, this default behavior is not correct on some machines, such as the
676 DEC Alpha, that store short integers in floating-point registers differently
677 than in integer registers. On those machines, the default widening will not
678 work correctly and you must define this macro to suppress that widening in
679 some cases. See the file `alpha.h' for details.
681 Do not define this macro if you do not define `SECONDARY_MEMORY_NEEDED' or
682 if widening MODE to a mode that is `BITS_PER_WORD' bits wide is correct for
684 /* #define SECONDARY_MEMORY_NEEDED_MODE(MODE) */
686 /* Normally the compiler avoids choosing registers that have been explicitly
687 mentioned in the rtl as spill registers (these registers are normally those
688 used to pass parameters and return values). However, some machines have so
689 few registers of certain classes that there would not be enough registers to
690 use as spill registers if this were done.
692 Define `SMALL_REGISTER_CLASSES' to be an expression with a non-zero value on
693 these machines. When this macro has a non-zero value, the compiler allows
694 registers explicitly used in the rtl to be used as spill registers but
695 avoids extending the lifetime of these registers.
697 It is always safe to define this macro with a non-zero value, but if you
698 unnecessarily define it, you will reduce the amount of optimizations that
699 can be performed in some cases. If you do not define this macro with a
700 non-zero value when it is required, the compiler will run out of spill
701 registers and print a fatal error message. For most machines, you should
702 not define this macro at all. */
703 /* #define SMALL_REGISTER_CLASSES */
705 /* A C expression whose value is nonzero if pseudos that have been assigned to
706 registers of class CLASS would likely be spilled because registers of CLASS
707 are needed for spill registers.
709 The default value of this macro returns 1 if CLASS has exactly one register
710 and zero otherwise. On most machines, this default should be used. Only
711 define this macro to some other expression if pseudo allocated by
712 `local-alloc.c' end up in memory because their hard registers were needed
713 for spill registers. If this macro returns nonzero for those classes, those
714 pseudos will only be allocated by `global.c', which knows how to reallocate
715 the pseudo to another register. If there would not be another register
716 available for reallocation, you should not change the definition of this
717 macro since the only effect of such a definition would be to slow down
718 register allocation. */
719 /* #define CLASS_LIKELY_SPILLED_P(CLASS) */
721 /* A C expression for the maximum number of consecutive registers of
722 class CLASS needed to hold a value of mode MODE.
724 This is closely related to the macro `HARD_REGNO_NREGS'. In fact, the value
725 of the macro `CLASS_MAX_NREGS (CLASS, MODE)' should be the maximum value of
726 `HARD_REGNO_NREGS (REGNO, MODE)' for all REGNO values in the class CLASS.
728 This macro helps control the handling of multiple-word values in
731 This declaration is required. */
732 #define CLASS_MAX_NREGS(CLASS, MODE) \
733 ((GET_MODE_SIZE (MODE) + UNITS_PER_WORD - 1) / UNITS_PER_WORD)
735 /* If defined, a C expression for a class that contains registers which the
736 compiler must always access in a mode that is the same size as the mode in
737 which it loaded the register.
739 For the example, loading 32-bit integer or floating-point objects into
740 floating-point registers on the Alpha extends them to 64-bits. Therefore
741 loading a 64-bit object and then storing it as a 32-bit object does not
742 store the low-order 32-bits, as would be the case for a normal register.
743 Therefore, `alpha.h' defines this macro as `FLOAT_REGS'. */
744 /* #define CLASS_CANNOT_CHANGE_SIZE */
746 /* A C expression that defines the machine-dependent operand constraint letters
747 (`I', `J', `K', .. 'P') that specify particular ranges of integer values.
748 If C is one of those letters, the expression should check that VALUE, an
749 integer, is in the appropriate range and return 1 if so, 0 otherwise. If C
750 is not one of those letters, the value should be 0 regardless of VALUE. */
751 #define CONST_OK_FOR_LETTER_P(VALUE, C) \
752 ( (C) == 'I' ? (VALUE) >= 0 && (VALUE) <= 3 \
753 : (C) == 'J' ? exact_log2 (VALUE) != -1 \
754 : (C) == 'K' ? exact_log2 (~(VALUE)) != -1 \
755 : (C) == 'L' ? (VALUE) >= 0 && (VALUE) <= 255 \
756 : (C) == 'M' ? (VALUE) >= -255 && (VALUE) <= 0 \
757 : (C) == 'N' ? (VALUE) >= -3 && (VALUE) <= 0 \
758 : (C) == 'O' ? (VALUE) >= 1 && (VALUE) <= 4 \
759 : (C) == 'P' ? (VALUE) >= -4 && (VALUE) <= -1 \
762 /* A C expression that defines the machine-dependent operand constraint letters
763 (`G', `H') that specify particular ranges of `const_double' values.
765 If C is one of those letters, the expression should check that VALUE, an RTX
766 of code `const_double', is in the appropriate range and return 1 if so, 0
767 otherwise. If C is not one of those letters, the value should be 0
770 `const_double' is used for all floating-point constants and for `DImode'
771 fixed-point constants. A given letter can accept either or both kinds of
772 values. It can use `GET_MODE' to distinguish between these kinds. */
773 #define CONST_DOUBLE_OK_FOR_LETTER_P(VALUE, C) 0
775 /* A C expression that defines the optional machine-dependent constraint
776 letters (`Q', `R', `S', `T', `U') that can be used to segregate specific
777 types of operands, usually memory references, for the target machine.
778 Normally this macro will not be defined. If it is required for a particular
779 target machine, it should return 1 if VALUE corresponds to the operand type
780 represented by the constraint letter C. If C is not defined as an extra
781 constraint, the value returned should be 0 regardless of VALUE.
783 For example, on the ROMP, load instructions cannot have their output in r0
784 if the memory reference contains a symbolic address. Constraint letter `Q'
785 is defined as representing a memory address that does *not* contain a
786 symbolic address. An alternative is specified with a `Q' constraint on the
787 input and `r' on the output. The next alternative specifies `m' on the
788 input and a register class that does not include r0 on the output. */
789 #define EXTRA_CONSTRAINT(VALUE, C) \
790 xstormy16_extra_constraint_p (VALUE, C)
793 /* Basic Stack Layout */
795 /* Define this macro if pushing a word onto the stack moves the stack pointer
796 to a smaller address.
798 When we say, "define this macro if ...," it means that the compiler checks
799 this macro only with `#ifdef' so the precise definition used does not
801 /* #define STACK_GROWS_DOWNWARD */
803 /* We want to use post-increment instructions to push things on the stack,
804 because we don't have any pre-increment ones. */
805 #define STACK_PUSH_CODE POST_INC
807 /* Define this macro if the addresses of local variable slots are at negative
808 offsets from the frame pointer. */
809 /* #define FRAME_GROWS_DOWNWARD */
811 /* Define this macro if successive arguments to a function occupy decreasing
812 addresses on the stack. */
813 #define ARGS_GROW_DOWNWARD 1
815 /* Offset from the frame pointer to the first local variable slot to be
818 If `FRAME_GROWS_DOWNWARD', find the next slot's offset by
819 subtracting the first slot's length from `STARTING_FRAME_OFFSET'.
820 Otherwise, it is found by adding the length of the first slot to
821 the value `STARTING_FRAME_OFFSET'. */
822 #define STARTING_FRAME_OFFSET 0
824 /* Offset from the stack pointer register to the first location at which
825 outgoing arguments are placed. If not specified, the default value of zero
826 is used. This is the proper value for most machines.
828 If `ARGS_GROW_DOWNWARD', this is the offset to the location above the first
829 location at which outgoing arguments are placed. */
830 /* #define STACK_POINTER_OFFSET */
832 /* Offset from the argument pointer register to the first argument's address.
833 On some machines it may depend on the data type of the function.
835 If `ARGS_GROW_DOWNWARD', this is the offset to the location above the first
836 argument's address. */
837 #define FIRST_PARM_OFFSET(FUNDECL) 0
839 /* Offset from the stack pointer register to an item dynamically allocated on
840 the stack, e.g., by `alloca'.
842 The default value for this macro is `STACK_POINTER_OFFSET' plus the length
843 of the outgoing arguments. The default is correct for most machines. See
844 `function.c' for details. */
845 /* #define STACK_DYNAMIC_OFFSET(FUNDECL) */
847 /* A C expression whose value is RTL representing the address in a stack frame
848 where the pointer to the caller's frame is stored. Assume that FRAMEADDR is
849 an RTL expression for the address of the stack frame itself.
851 If you don't define this macro, the default is to return the value of
852 FRAMEADDR--that is, the stack frame address is also the address of the stack
853 word that points to the previous frame. */
854 /* #define DYNAMIC_CHAIN_ADDRESS(FRAMEADDR) */
856 /* If defined, a C expression that produces the machine-specific code to setup
857 the stack so that arbitrary frames can be accessed. For example, on the
858 Sparc, we must flush all of the register windows to the stack before we can
859 access arbitrary stack frames. This macro will seldom need to be defined. */
860 /* #define SETUP_FRAME_ADDRESSES() */
862 /* A C expression whose value is RTL representing the value of the return
863 address for the frame COUNT steps up from the current frame, after the
864 prologue. FRAMEADDR is the frame pointer of the COUNT frame, or the frame
865 pointer of the COUNT - 1 frame if `RETURN_ADDR_IN_PREVIOUS_FRAME' is
868 The value of the expression must always be the correct address when COUNT is
869 zero, but may be `NULL_RTX' if there is not way to determine the return
870 address of other frames. */
871 #define RETURN_ADDR_RTX(COUNT, FRAMEADDR) \
873 ? gen_rtx_MEM (Pmode, arg_pointer_rtx) \
876 /* Define this if the return address of a particular stack frame is
877 accessed from the frame pointer of the previous stack frame. */
878 /* #define RETURN_ADDR_IN_PREVIOUS_FRAME */
880 /* A C expression whose value is RTL representing the location of the incoming
881 return address at the beginning of any function, before the prologue. This
882 RTL is either a `REG', indicating that the return value is saved in `REG',
883 or a `MEM' representing a location in the stack.
885 You only need to define this macro if you want to support call frame
886 debugging information like that provided by DWARF 2. */
887 #define INCOMING_RETURN_ADDR_RTX \
888 gen_rtx_MEM (SImode, gen_rtx_PLUS (Pmode, stack_pointer_rtx, GEN_INT (-4)))
890 /* A C expression whose value is an integer giving the offset, in bytes, from
891 the value of the stack pointer register to the top of the stack frame at the
892 beginning of any function, before the prologue. The top of the frame is
893 defined to be the value of the stack pointer in the previous frame, just
894 before the call instruction.
896 You only need to define this macro if you want to support call frame
897 debugging information like that provided by DWARF 2. */
898 #define INCOMING_FRAME_SP_OFFSET (xstormy16_interrupt_function_p () ? 6 : 4)
901 /* Stack Checking. */
903 /* A nonzero value if stack checking is done by the configuration files in a
904 machine-dependent manner. You should define this macro if stack checking is
905 require by the ABI of your machine or if you would like to have to stack
906 checking in some more efficient way than GNU CC's portable approach. The
907 default value of this macro is zero. */
908 /* #define STACK_CHECK_BUILTIN */
910 /* An integer representing the interval at which GNU CC must generate stack
911 probe instructions. You will normally define this macro to be no larger
912 than the size of the "guard pages" at the end of a stack area. The default
913 value of 4096 is suitable for most systems. */
914 /* #define STACK_CHECK_PROBE_INTERVAL */
916 /* A integer which is nonzero if GNU CC should perform the stack probe as a
917 load instruction and zero if GNU CC should use a store instruction. The
918 default is zero, which is the most efficient choice on most systems. */
919 /* #define STACK_CHECK_PROBE_LOAD */
921 /* The number of bytes of stack needed to recover from a stack overflow, for
922 languages where such a recovery is supported. The default value of 75 words
923 should be adequate for most machines. */
924 /* #define STACK_CHECK_PROTECT */
926 /* The maximum size of a stack frame, in bytes. GNU CC will generate probe
927 instructions in non-leaf functions to ensure at least this many bytes of
928 stack are available. If a stack frame is larger than this size, stack
929 checking will not be reliable and GNU CC will issue a warning. The default
930 is chosen so that GNU CC only generates one instruction on most systems.
931 You should normally not change the default value of this macro. */
932 /* #define STACK_CHECK_MAX_FRAME_SIZE */
934 /* GNU CC uses this value to generate the above warning message. It represents
935 the amount of fixed frame used by a function, not including space for any
936 callee-saved registers, temporaries and user variables. You need only
937 specify an upper bound for this amount and will normally use the default of
939 /* #define STACK_CHECK_FIXED_FRAME_SIZE */
941 /* The maximum size, in bytes, of an object that GNU CC will place in the fixed
942 area of the stack frame when the user specifies `-fstack-check'. GNU CC
943 computed the default from the values of the above macros and you will
944 normally not need to override that default. */
945 /* #define STACK_CHECK_MAX_VAR_SIZE */
948 /* Register That Address the Stack Frame. */
950 /* The register number of the stack pointer register, which must also be a
951 fixed register according to `FIXED_REGISTERS'. On most machines, the
952 hardware determines which register this is. */
953 #define STACK_POINTER_REGNUM 15
955 /* The register number of the frame pointer register, which is used to access
956 automatic variables in the stack frame. On some machines, the hardware
957 determines which register this is. On other machines, you can choose any
958 register you wish for this purpose. */
959 #define FRAME_POINTER_REGNUM 17
961 /* On some machines the offset between the frame pointer and starting offset of
962 the automatic variables is not known until after register allocation has
963 been done (for example, because the saved registers are between these two
964 locations). On those machines, define `FRAME_POINTER_REGNUM' the number of
965 a special, fixed register to be used internally until the offset is known,
966 and define `HARD_FRAME_POINTER_REGNUM' to be actual the hard register number
967 used for the frame pointer.
969 You should define this macro only in the very rare circumstances when it is
970 not possible to calculate the offset between the frame pointer and the
971 automatic variables until after register allocation has been completed.
972 When this macro is defined, you must also indicate in your definition of
973 `ELIMINABLE_REGS' how to eliminate `FRAME_POINTER_REGNUM' into either
974 `HARD_FRAME_POINTER_REGNUM' or `STACK_POINTER_REGNUM'.
976 Do not define this macro if it would be the same as `FRAME_POINTER_REGNUM'. */
977 #define HARD_FRAME_POINTER_REGNUM 13
979 /* The register number of the arg pointer register, which is used to access the
980 function's argument list. On some machines, this is the same as the frame
981 pointer register. On some machines, the hardware determines which register
982 this is. On other machines, you can choose any register you wish for this
983 purpose. If this is not the same register as the frame pointer register,
984 then you must mark it as a fixed register according to `FIXED_REGISTERS', or
985 arrange to be able to eliminate it. */
986 #define ARG_POINTER_REGNUM 18
988 /* The register number of the return address pointer register, which is used to
989 access the current function's return address from the stack. On some
990 machines, the return address is not at a fixed offset from the frame pointer
991 or stack pointer or argument pointer. This register can be defined to point
992 to the return address on the stack, and then be converted by
993 `ELIMINABLE_REGS' into either the frame pointer or stack pointer.
995 Do not define this macro unless there is no other way to get the return
996 address from the stack. */
997 /* #define RETURN_ADDRESS_POINTER_REGNUM */
999 /* Register numbers used for passing a function's static chain pointer. If
1000 register windows are used, the register number as seen by the called
1001 function is `STATIC_CHAIN_INCOMING_REGNUM', while the register number as
1002 seen by the calling function is `STATIC_CHAIN_REGNUM'. If these registers
1003 are the same, `STATIC_CHAIN_INCOMING_REGNUM' need not be defined.
1005 The static chain register need not be a fixed register.
1007 If the static chain is passed in memory, these macros should not be defined;
1008 instead, the next two macros should be defined. */
1009 #define STATIC_CHAIN_REGNUM 1
1010 /* #define STATIC_CHAIN_INCOMING_REGNUM */
1012 /* If the static chain is passed in memory, these macros provide rtx giving
1013 `mem' expressions that denote where they are stored. `STATIC_CHAIN' and
1014 `STATIC_CHAIN_INCOMING' give the locations as seen by the calling and called
1015 functions, respectively. Often the former will be at an offset from the
1016 stack pointer and the latter at an offset from the frame pointer.
1018 The variables `stack_pointer_rtx', `frame_pointer_rtx', and
1019 `arg_pointer_rtx' will have been initialized prior to the use of these
1020 macros and should be used to refer to those items.
1022 If the static chain is passed in a register, the two previous
1023 macros should be defined instead. */
1024 /* #define STATIC_CHAIN */
1025 /* #define STATIC_CHAIN_INCOMING */
1028 /* Eliminating the Frame Pointer and the Arg Pointer */
1030 /* A C expression which is nonzero if a function must have and use a frame
1031 pointer. This expression is evaluated in the reload pass. If its value is
1032 nonzero the function will have a frame pointer.
1034 The expression can in principle examine the current function and decide
1035 according to the facts, but on most machines the constant 0 or the constant
1036 1 suffices. Use 0 when the machine allows code to be generated with no
1037 frame pointer, and doing so saves some time or space. Use 1 when there is
1038 no possible advantage to avoiding a frame pointer.
1040 In certain cases, the compiler does not know how to produce valid code
1041 without a frame pointer. The compiler recognizes those cases and
1042 automatically gives the function a frame pointer regardless of what
1043 `FRAME_POINTER_REQUIRED' says. You don't need to worry about them.
1045 In a function that does not require a frame pointer, the frame pointer
1046 register can be allocated for ordinary usage, unless you mark it as a fixed
1047 register. See `FIXED_REGISTERS' for more information. */
1048 #define FRAME_POINTER_REQUIRED 0
1050 /* A C statement to store in the variable DEPTH_VAR the difference between the
1051 frame pointer and the stack pointer values immediately after the function
1052 prologue. The value would be computed from information such as the result
1053 of `get_frame_size ()' and the tables of registers `regs_ever_live' and
1056 If `ELIMINABLE_REGS' is defined, this macro will be not be used and need not
1057 be defined. Otherwise, it must be defined even if `FRAME_POINTER_REQUIRED'
1058 is defined to always be true; in that case, you may set DEPTH_VAR to
1060 /* #define INITIAL_FRAME_POINTER_OFFSET(DEPTH_VAR) */
1062 /* If defined, this macro specifies a table of register pairs used to eliminate
1063 unneeded registers that point into the stack frame. If it is not defined,
1064 the only elimination attempted by the compiler is to replace references to
1065 the frame pointer with references to the stack pointer.
1067 The definition of this macro is a list of structure initializations, each of
1068 which specifies an original and replacement register.
1071 #define ELIMINABLE_REGS \
1073 {FRAME_POINTER_REGNUM, STACK_POINTER_REGNUM}, \
1074 {FRAME_POINTER_REGNUM, HARD_FRAME_POINTER_REGNUM}, \
1075 {ARG_POINTER_REGNUM, STACK_POINTER_REGNUM}, \
1076 {ARG_POINTER_REGNUM, HARD_FRAME_POINTER_REGNUM}, \
1079 /* A C expression that returns non-zero if the compiler is allowed to try to
1080 replace register number FROM with register number TO. This macro need only
1081 be defined if `ELIMINABLE_REGS' is defined, and will usually be the constant
1082 1, since most of the cases preventing register elimination are things that
1083 the compiler already knows about. */
1085 #define CAN_ELIMINATE(FROM, TO) \
1086 ((FROM) == ARG_POINTER_REGNUM && (TO) == STACK_POINTER_REGNUM \
1087 ? ! frame_pointer_needed \
1090 /* This macro is similar to `INITIAL_FRAME_POINTER_OFFSET'. It specifies the
1091 initial difference between the specified pair of registers. This macro must
1092 be defined if `ELIMINABLE_REGS' is defined. */
1093 #define INITIAL_ELIMINATION_OFFSET(FROM, TO, OFFSET) \
1094 (OFFSET) = xstormy16_initial_elimination_offset (FROM, TO)
1097 /* Passing Function Arguments on the Stack */
1099 /* Define this macro if an argument declared in a prototype as an integral type
1100 smaller than `int' should actually be passed as an `int'. In addition to
1101 avoiding errors in certain cases of mismatch, it also makes for better code
1102 on certain machines. */
1103 #define PROMOTE_PROTOTYPES 1
1105 /* A C expression that is the number of bytes actually pushed onto the stack
1106 when an instruction attempts to push NPUSHED bytes.
1108 If the target machine does not have a push instruction, do not define this
1109 macro. That directs GNU CC to use an alternate strategy: to allocate the
1110 entire argument block and then store the arguments into it.
1112 On some machines, the definition
1114 #define PUSH_ROUNDING(BYTES) (BYTES)
1116 will suffice. But on other machines, instructions that appear to push one
1117 byte actually push two bytes in an attempt to maintain alignment. Then the
1118 definition should be
1120 #define PUSH_ROUNDING(BYTES) (((BYTES) + 1) & ~1) */
1121 #define PUSH_ROUNDING(BYTES) (((BYTES) + 1) & ~1)
1123 /* If defined, the maximum amount of space required for outgoing arguments will
1124 be computed and placed into the variable
1125 `current_function_outgoing_args_size'. No space will be pushed onto the
1126 stack for each call; instead, the function prologue should increase the
1127 stack frame size by this amount.
1129 Defining both `PUSH_ROUNDING' and `ACCUMULATE_OUTGOING_ARGS' is not
1131 /* #define ACCUMULATE_OUTGOING_ARGS */
1133 /* Define this macro if functions should assume that stack space has been
1134 allocated for arguments even when their values are passed in registers.
1136 The value of this macro is the size, in bytes, of the area reserved for
1137 arguments passed in registers for the function represented by FNDECL.
1139 This space can be allocated by the caller, or be a part of the
1140 machine-dependent stack frame: `OUTGOING_REG_PARM_STACK_SPACE' says
1142 /* #define REG_PARM_STACK_SPACE(FNDECL) */
1144 /* Define these macros in addition to the one above if functions might allocate
1145 stack space for arguments even when their values are passed in registers.
1146 These should be used when the stack space allocated for arguments in
1147 registers is not a simple constant independent of the function declaration.
1149 The value of the first macro is the size, in bytes, of the area that we
1150 should initially assume would be reserved for arguments passed in registers.
1152 The value of the second macro is the actual size, in bytes, of the area that
1153 will be reserved for arguments passed in registers. This takes two
1154 arguments: an integer representing the number of bytes of fixed sized
1155 arguments on the stack, and a tree representing the number of bytes of
1156 variable sized arguments on the stack.
1158 When these macros are defined, `REG_PARM_STACK_SPACE' will only be called
1159 for libcall functions, the current function, or for a function being called
1160 when it is known that such stack space must be allocated. In each case this
1161 value can be easily computed.
1163 When deciding whether a called function needs such stack space, and how much
1164 space to reserve, GNU CC uses these two macros instead of
1165 `REG_PARM_STACK_SPACE'. */
1166 /* #define MAYBE_REG_PARM_STACK_SPACE */
1167 /* #define FINAL_REG_PARM_STACK_SPACE(CONST_SIZE, VAR_SIZE) */
1169 /* Define this if it is the responsibility of the caller to allocate the area
1170 reserved for arguments passed in registers.
1172 If `ACCUMULATE_OUTGOING_ARGS' is defined, this macro controls whether the
1173 space for these arguments counts in the value of
1174 `current_function_outgoing_args_size'. */
1175 /* #define OUTGOING_REG_PARM_STACK_SPACE */
1177 /* Define this macro if `REG_PARM_STACK_SPACE' is defined, but the stack
1178 parameters don't skip the area specified by it.
1180 Normally, when a parameter is not passed in registers, it is placed on the
1181 stack beyond the `REG_PARM_STACK_SPACE' area. Defining this macro
1182 suppresses this behavior and causes the parameter to be passed on the stack
1183 in its natural location. */
1184 /* #define STACK_PARMS_IN_REG_PARM_AREA */
1186 /* A C expression that should indicate the number of bytes of its own arguments
1187 that a function pops on returning, or 0 if the function pops no arguments
1188 and the caller must therefore pop them all after the function returns.
1190 FUNDECL is a C variable whose value is a tree node that describes the
1191 function in question. Normally it is a node of type `FUNCTION_DECL' that
1192 describes the declaration of the function. From this it is possible to
1193 obtain the DECL_ATTRIBUTES of the function.
1195 FUNTYPE is a C variable whose value is a tree node that describes the
1196 function in question. Normally it is a node of type `FUNCTION_TYPE' that
1197 describes the data type of the function. From this it is possible to obtain
1198 the data types of the value and arguments (if known).
1200 When a call to a library function is being considered, FUNTYPE will contain
1201 an identifier node for the library function. Thus, if you need to
1202 distinguish among various library functions, you can do so by their names.
1203 Note that "library function" in this context means a function used to
1204 perform arithmetic, whose name is known specially in the compiler and was
1205 not mentioned in the C code being compiled.
1207 STACK-SIZE is the number of bytes of arguments passed on the stack. If a
1208 variable number of bytes is passed, it is zero, and argument popping will
1209 always be the responsibility of the calling function.
1211 On the Vax, all functions always pop their arguments, so the definition of
1212 this macro is STACK-SIZE. On the 68000, using the standard calling
1213 convention, no functions pop their arguments, so the value of the macro is
1214 always 0 in this case. But an alternative calling convention is available
1215 in which functions that take a fixed number of arguments pop them but other
1216 functions (such as `printf') pop nothing (the caller pops all). When this
1217 convention is in use, FUNTYPE is examined to determine whether a function
1218 takes a fixed number of arguments. */
1219 #define RETURN_POPS_ARGS(FUNDECL, FUNTYPE, STACK_SIZE) 0
1222 /* Function Arguments in Registers */
1224 #define NUM_ARGUMENT_REGISTERS 6
1225 #define FIRST_ARGUMENT_REGISTER 2
1227 #define XSTORMY16_WORD_SIZE(TYPE, MODE) \
1228 ((((TYPE) ? int_size_in_bytes (TYPE) : GET_MODE_SIZE (MODE)) \
1232 /* A C expression that controls whether a function argument is passed in a
1233 register, and which register.
1235 The arguments are CUM, of type CUMULATIVE_ARGS, which summarizes
1236 (in a way defined by INIT_CUMULATIVE_ARGS and FUNCTION_ARG_ADVANCE)
1237 all of the previous arguments so far passed in registers; MODE, the
1238 machine mode of the argument; TYPE, the data type of the argument
1239 as a tree node or 0 if that is not known (which happens for C
1240 support library functions); and NAMED, which is 1 for an ordinary
1241 argument and 0 for nameless arguments that correspond to `...' in
1242 the called function's prototype.
1244 The value of the expression should either be a `reg' RTX for the hard
1245 register in which to pass the argument, or zero to pass the argument on the
1248 For machines like the Vax and 68000, where normally all arguments are
1249 pushed, zero suffices as a definition.
1251 The usual way to make the ANSI library `stdarg.h' work on a machine where
1252 some arguments are usually passed in registers, is to cause nameless
1253 arguments to be passed on the stack instead. This is done by making
1254 `FUNCTION_ARG' return 0 whenever NAMED is 0.
1256 You may use the macro `MUST_PASS_IN_STACK (MODE, TYPE)' in the definition of
1257 this macro to determine if this argument is of a type that must be passed in
1258 the stack. If `REG_PARM_STACK_SPACE' is not defined and `FUNCTION_ARG'
1259 returns non-zero for such an argument, the compiler will abort. If
1260 `REG_PARM_STACK_SPACE' is defined, the argument will be computed in the
1261 stack and then loaded into a register. */
1262 #define FUNCTION_ARG(CUM, MODE, TYPE, NAMED) \
1263 ((MODE) == VOIDmode ? const0_rtx \
1264 : (CUM) + XSTORMY16_WORD_SIZE (TYPE, MODE) > NUM_ARGUMENT_REGISTERS ? 0 \
1265 : gen_rtx_REG (MODE, (CUM) + 2))
1267 /* Define this macro if the target machine has "register windows", so that the
1268 register in which a function sees an arguments is not necessarily the same
1269 as the one in which the caller passed the argument.
1271 For such machines, `FUNCTION_ARG' computes the register in which the caller
1272 passes the value, and `FUNCTION_INCOMING_ARG' should be defined in a similar
1273 fashion to tell the function being called where the arguments will arrive.
1275 If `FUNCTION_INCOMING_ARG' is not defined, `FUNCTION_ARG' serves both
1277 /* #define FUNCTION_INCOMING_ARG(CUM, MODE, TYPE, NAMED) */
1279 /* A C expression for the number of words, at the beginning of an argument,
1280 must be put in registers. The value must be zero for arguments that are
1281 passed entirely in registers or that are entirely pushed on the stack.
1283 On some machines, certain arguments must be passed partially in registers
1284 and partially in memory. On these machines, typically the first N words of
1285 arguments are passed in registers, and the rest on the stack. If a
1286 multi-word argument (a `double' or a structure) crosses that boundary, its
1287 first few words must be passed in registers and the rest must be pushed.
1288 This macro tells the compiler when this occurs, and how many of the words
1289 should go in registers.
1291 `FUNCTION_ARG' for these arguments should return the first register to be
1292 used by the caller for this argument; likewise `FUNCTION_INCOMING_ARG', for
1293 the called function. */
1294 #define FUNCTION_ARG_PARTIAL_NREGS(CUM, MODE, TYPE, NAMED) 0
1296 /* A C expression that indicates when an argument must be passed by reference.
1297 If nonzero for an argument, a copy of that argument is made in memory and a
1298 pointer to the argument is passed instead of the argument itself. The
1299 pointer is passed in whatever way is appropriate for passing a pointer to
1302 On machines where `REG_PARM_STACK_SPACE' is not defined, a suitable
1303 definition of this macro might be
1304 #define FUNCTION_ARG_PASS_BY_REFERENCE(CUM, MODE, TYPE, NAMED) \
1305 MUST_PASS_IN_STACK (MODE, TYPE) */
1306 #define FUNCTION_ARG_PASS_BY_REFERENCE(CUM, MODE, TYPE, NAMED) 0
1308 /* If defined, a C expression that indicates when it is more
1309 desirable to keep an argument passed by invisible reference as a
1310 reference, rather than copying it to a pseudo register. */
1311 /* #define FUNCTION_ARG_KEEP_AS_REFERENCE(CUM, MODE, TYPE, NAMED) */
1313 /* If defined, a C expression that indicates when it is the called function's
1314 responsibility to make a copy of arguments passed by invisible reference.
1315 Normally, the caller makes a copy and passes the address of the copy to the
1316 routine being called. When FUNCTION_ARG_CALLEE_COPIES is defined and is
1317 nonzero, the caller does not make a copy. Instead, it passes a pointer to
1318 the "live" value. The called function must not modify this value. If it
1319 can be determined that the value won't be modified, it need not make a copy;
1320 otherwise a copy must be made. */
1321 /* #define FUNCTION_ARG_CALLEE_COPIES(CUM, MODE, TYPE, NAMED) */
1323 /* A C type for declaring a variable that is used as the first argument of
1324 `FUNCTION_ARG' and other related values. For some target machines, the type
1325 `int' suffices and can hold the number of bytes of argument so far.
1327 There is no need to record in `CUMULATIVE_ARGS' anything about the arguments
1328 that have been passed on the stack. The compiler has other variables to
1329 keep track of that. For target machines on which all arguments are passed
1330 on the stack, there is no need to store anything in `CUMULATIVE_ARGS';
1331 however, the data structure must exist and should not be empty, so use
1334 For this platform, the value of CUMULATIVE_ARGS is the number of words
1335 of arguments that have been passed in registers so far. */
1336 typedef int CUMULATIVE_ARGS;
1338 /* A C statement (sans semicolon) for initializing the variable CUM for the
1339 state at the beginning of the argument list. The variable has type
1340 `CUMULATIVE_ARGS'. The value of FNTYPE is the tree node for the data type
1341 of the function which will receive the args, or 0 if the args are to a
1342 compiler support library function. The value of INDIRECT is nonzero when
1343 processing an indirect call, for example a call through a function pointer.
1344 The value of INDIRECT is zero for a call to an explicitly named function, a
1345 library function call, or when `INIT_CUMULATIVE_ARGS' is used to find
1346 arguments for the function being compiled.
1348 When processing a call to a compiler support library function, LIBNAME
1349 identifies which one. It is a `symbol_ref' rtx which contains the name of
1350 the function, as a string. LIBNAME is 0 when an ordinary C function call is
1351 being processed. Thus, each time this macro is called, either LIBNAME or
1352 FNTYPE is nonzero, but never both of them at once. */
1353 #define INIT_CUMULATIVE_ARGS(CUM, FNTYPE, LIBNAME, INDIRECT) (CUM) = 0
1355 /* Like `INIT_CUMULATIVE_ARGS' but overrides it for the purposes of finding the
1356 arguments for the function being compiled. If this macro is undefined,
1357 `INIT_CUMULATIVE_ARGS' is used instead.
1359 The value passed for LIBNAME is always 0, since library routines with
1360 special calling conventions are never compiled with GNU CC. The argument
1361 LIBNAME exists for symmetry with `INIT_CUMULATIVE_ARGS'. */
1362 /* #define INIT_CUMULATIVE_INCOMING_ARGS(CUM, FNTYPE, LIBNAME) */
1364 /* A C statement (sans semicolon) to update the summarizer variable CUM to
1365 advance past an argument in the argument list. The values MODE, TYPE and
1366 NAMED describe that argument. Once this is done, the variable CUM is
1367 suitable for analyzing the *following* argument with `FUNCTION_ARG', etc.
1369 This macro need not do anything if the argument in question was passed on
1370 the stack. The compiler knows how to track the amount of stack space used
1371 for arguments without any special help. */
1372 #define FUNCTION_ARG_ADVANCE(CUM, MODE, TYPE, NAMED) \
1373 ((CUM) = xstormy16_function_arg_advance (CUM, MODE, TYPE, NAMED))
1375 /* If defined, a C expression which determines whether, and in which direction,
1376 to pad out an argument with extra space. The value should be of type `enum
1377 direction': either `upward' to pad above the argument, `downward' to pad
1378 below, or `none' to inhibit padding.
1380 The *amount* of padding is always just enough to reach the next multiple of
1381 `FUNCTION_ARG_BOUNDARY'; this macro does not control it.
1383 This macro has a default definition which is right for most systems. For
1384 little-endian machines, the default is to pad upward. For big-endian
1385 machines, the default is to pad downward for an argument of constant size
1386 shorter than an `int', and upward otherwise. */
1387 /* #define FUNCTION_ARG_PADDING(MODE, TYPE) */
1389 /* If defined, a C expression that gives the alignment boundary, in bits, of an
1390 argument with the specified mode and type. If it is not defined,
1391 `PARM_BOUNDARY' is used for all arguments. */
1392 /* #define FUNCTION_ARG_BOUNDARY(MODE, TYPE) */
1394 /* A C expression that is nonzero if REGNO is the number of a hard register in
1395 which function arguments are sometimes passed. This does *not* include
1396 implicit arguments such as the static chain and the structure-value address.
1397 On many machines, no registers can be used for this purpose since all
1398 function arguments are pushed on the stack. */
1399 #define FUNCTION_ARG_REGNO_P(REGNO) \
1400 ((REGNO) >= FIRST_ARGUMENT_REGISTER \
1401 && (REGNO) < FIRST_ARGUMENT_REGISTER + NUM_ARGUMENT_REGISTERS)
1404 /* How Scalar Function Values are Returned */
1406 /* The number of the hard register that is used to return a scalar value from a
1408 #define RETURN_VALUE_REGNUM FIRST_ARGUMENT_REGISTER
1410 /* A C expression to create an RTX representing the place where a function
1411 returns a value of data type VALTYPE. VALTYPE is a tree node representing a
1412 data type. Write `TYPE_MODE (VALTYPE)' to get the machine mode used to
1413 represent that type. On many machines, only the mode is relevant.
1414 (Actually, on most machines, scalar values are returned in the same place
1415 regardless of mode).
1417 If `PROMOTE_FUNCTION_RETURN' is defined, you must apply the same promotion
1418 rules specified in `PROMOTE_MODE' if VALTYPE is a scalar type.
1420 If the precise function being called is known, FUNC is a tree node
1421 (`FUNCTION_DECL') for it; otherwise, FUNC is a null pointer. This makes it
1422 possible to use a different value-returning convention for specific
1423 functions when all their calls are known.
1425 `FUNCTION_VALUE' is not used for return vales with aggregate data types,
1426 because these are returned in another way. See `STRUCT_VALUE_REGNUM' and
1427 related macros, below. */
1428 #define FUNCTION_VALUE(VALTYPE, FUNC) \
1429 xstormy16_function_value (VALTYPE, FUNC)
1432 /* Define this macro if the target machine has "register windows" so that the
1433 register in which a function returns its value is not the same as the one in
1434 which the caller sees the value.
1436 For such machines, `FUNCTION_VALUE' computes the register in which the
1437 caller will see the value. `FUNCTION_OUTGOING_VALUE' should be defined in a
1438 similar fashion to tell the function where to put the value.
1440 If `FUNCTION_OUTGOING_VALUE' is not defined, `FUNCTION_VALUE' serves both
1443 `FUNCTION_OUTGOING_VALUE' is not used for return vales with aggregate data
1444 types, because these are returned in another way. See `STRUCT_VALUE_REGNUM'
1445 and related macros, below. */
1446 /* #define FUNCTION_OUTGOING_VALUE(VALTYPE, FUNC) */
1448 /* A C expression to create an RTX representing the place where a library
1449 function returns a value of mode MODE.
1451 Note that "library function" in this context means a compiler support
1452 routine, used to perform arithmetic, whose name is known specially by the
1453 compiler and was not mentioned in the C code being compiled.
1455 The definition of `LIBRARY_VALUE' need not be concerned aggregate data
1456 types, because none of the library functions returns such types. */
1457 #define LIBCALL_VALUE(MODE) gen_rtx_REG (MODE, RETURN_VALUE_REGNUM)
1459 /* A C expression that is nonzero if REGNO is the number of a hard register in
1460 which the values of called function may come back.
1462 A register whose use for returning values is limited to serving as the
1463 second of a pair (for a value of type `double', say) need not be recognized
1464 by this macro. So for most machines, this definition suffices:
1466 #define FUNCTION_VALUE_REGNO_P(N) ((N) == RETURN)
1468 If the machine has register windows, so that the caller and the called
1469 function use different registers for the return value, this macro should
1470 recognize only the caller's register numbers. */
1471 #define FUNCTION_VALUE_REGNO_P(REGNO) ((REGNO) == RETURN_VALUE_REGNUM)
1473 /* Define this macro if `untyped_call' and `untyped_return' need more space
1474 than is implied by `FUNCTION_VALUE_REGNO_P' for saving and restoring an
1475 arbitrary return value. */
1476 /* #define APPLY_RESULT_SIZE */
1479 /* How Large Values are Returned */
1481 /* A C expression which can inhibit the returning of certain function values in
1482 registers, based on the type of value. A nonzero value says to return the
1483 function value in memory, just as large structures are always returned.
1484 Here TYPE will be a C expression of type `tree', representing the data type
1487 Note that values of mode `BLKmode' must be explicitly handled by this macro.
1488 Also, the option `-fpcc-struct-return' takes effect regardless of this
1489 macro. On most systems, it is possible to leave the macro undefined; this
1490 causes a default definition to be used, whose value is the constant 1 for
1491 `BLKmode' values, and 0 otherwise.
1493 Do not use this macro to indicate that structures and unions should always
1494 be returned in memory. You should instead use `DEFAULT_PCC_STRUCT_RETURN'
1495 to indicate this. */
1496 #define RETURN_IN_MEMORY(TYPE) \
1497 (int_size_in_bytes (TYPE) > UNITS_PER_WORD * NUM_ARGUMENT_REGISTERS)
1499 /* Define this macro to be 1 if all structure and union return values must be
1500 in memory. Since this results in slower code, this should be defined only
1501 if needed for compatibility with other compilers or with an ABI. If you
1502 define this macro to be 0, then the conventions used for structure and union
1503 return values are decided by the `RETURN_IN_MEMORY' macro.
1505 If not defined, this defaults to the value 1. */
1506 /* #define DEFAULT_PCC_STRUCT_RETURN 0 */
1508 /* If the structure value address is passed in a register, then
1509 `STRUCT_VALUE_REGNUM' should be the number of that register. */
1510 /* #define STRUCT_VALUE_REGNUM */
1512 /* If the structure value address is not passed in a register, define
1513 `STRUCT_VALUE' as an expression returning an RTX for the place where the
1514 address is passed. If it returns 0, the address is passed as an "invisible"
1516 #define STRUCT_VALUE 0
1518 /* On some architectures the place where the structure value address is found
1519 by the called function is not the same place that the caller put it. This
1520 can be due to register windows, or it could be because the function prologue
1521 moves it to a different place.
1523 If the incoming location of the structure value address is in a register,
1524 define this macro as the register number. */
1525 /* #define STRUCT_VALUE_INCOMING_REGNUM */
1527 /* If the incoming location is not a register, then you should define
1528 `STRUCT_VALUE_INCOMING' as an expression for an RTX for where the called
1529 function should find the value. If it should find the value on the stack,
1530 define this to create a `mem' which refers to the frame pointer. A
1531 definition of 0 means that the address is passed as an "invisible" first
1533 /* #define STRUCT_VALUE_INCOMING */
1535 /* Define this macro if the usual system convention on the target machine for
1536 returning structures and unions is for the called function to return the
1537 address of a static variable containing the value.
1539 Do not define this if the usual system convention is for the caller to pass
1540 an address to the subroutine.
1542 This macro has effect in `-fpcc-struct-return' mode, but it does nothing
1543 when you use `-freg-struct-return' mode. */
1544 /* #define PCC_STATIC_STRUCT_RETURN */
1547 /* Caller-Saves Register Allocation */
1549 /* Define this macro if function calls on the target machine do not preserve
1550 any registers; in other words, if `CALL_USED_REGISTERS' has 1 for all
1551 registers. This macro enables `-fcaller-saves' by default. Eventually that
1552 option will be enabled by default on all machines and both the option and
1553 this macro will be eliminated. */
1554 /* #define DEFAULT_CALLER_SAVES */
1556 /* A C expression to determine whether it is worthwhile to consider placing a
1557 pseudo-register in a call-clobbered hard register and saving and restoring
1558 it around each function call. The expression should be 1 when this is worth
1559 doing, and 0 otherwise.
1561 If you don't define this macro, a default is used which is good on most
1562 machines: `4 * CALLS < REFS'. */
1563 /* #define CALLER_SAVE_PROFITABLE(REFS, CALLS) */
1566 /* Function Entry and Exit */
1568 /* Define this macro as a C expression that is nonzero if the return
1569 instruction or the function epilogue ignores the value of the stack pointer;
1570 in other words, if it is safe to delete an instruction to adjust the stack
1571 pointer before a return from the function.
1573 Note that this macro's value is relevant only for functions for which frame
1574 pointers are maintained. It is never safe to delete a final stack
1575 adjustment in a function that has no frame pointer, and the compiler knows
1576 this regardless of `EXIT_IGNORE_STACK'. */
1577 /* #define EXIT_IGNORE_STACK */
1579 /* Define this macro as a C expression that is nonzero for registers
1580 are used by the epilogue or the `return' pattern. The stack and
1581 frame pointer registers are already be assumed to be used as
1583 #define EPILOGUE_USES(REGNO) \
1584 xstormy16_epilogue_uses (REGNO)
1586 /* Define this macro if the function epilogue contains delay slots to which
1587 instructions from the rest of the function can be "moved". The definition
1588 should be a C expression whose value is an integer representing the number
1589 of delay slots there. */
1590 /* #define DELAY_SLOTS_FOR_EPILOGUE */
1592 /* A C expression that returns 1 if INSN can be placed in delay slot number N
1595 The argument N is an integer which identifies the delay slot now being
1596 considered (since different slots may have different rules of eligibility).
1597 It is never negative and is always less than the number of epilogue delay
1598 slots (what `DELAY_SLOTS_FOR_EPILOGUE' returns). If you reject a particular
1599 insn for a given delay slot, in principle, it may be reconsidered for a
1600 subsequent delay slot. Also, other insns may (at least in principle) be
1601 considered for the so far unfilled delay slot.
1603 The insns accepted to fill the epilogue delay slots are put in an
1604 RTL list made with `insn_list' objects, stored in the variable
1605 `current_function_epilogue_delay_list'. The insn for the first
1606 delay slot comes first in the list. Your definition of the macro
1607 `FUNCTION_EPILOGUE' should fill the delay slots by outputting the
1608 insns in this list, usually by calling `final_scan_insn'.
1610 You need not define this macro if you did not define
1611 `DELAY_SLOTS_FOR_EPILOGUE'. */
1612 /* #define ELIGIBLE_FOR_EPILOGUE_DELAY(INSN, N) */
1614 /* A C compound statement that outputs the assembler code for a thunk function,
1615 used to implement C++ virtual function calls with multiple inheritance. The
1616 thunk acts as a wrapper around a virtual function, adjusting the implicit
1617 object parameter before handing control off to the real function.
1619 First, emit code to add the integer DELTA to the location that contains the
1620 incoming first argument. Assume that this argument contains a pointer, and
1621 is the one used to pass the `this' pointer in C++. This is the incoming
1622 argument *before* the function prologue, e.g. `%o0' on a sparc. The
1623 addition must preserve the values of all other incoming arguments.
1625 After the addition, emit code to jump to FUNCTION, which is a
1626 `FUNCTION_DECL'. This is a direct pure jump, not a call, and does not touch
1627 the return address. Hence returning from FUNCTION will return to whoever
1628 called the current `thunk'.
1630 The effect must be as if @var{function} had been called directly
1631 with the adjusted first argument. This macro is responsible for
1632 emitting all of the code for a thunk function;
1633 TARGET_ASM_FUNCTION_PROLOGUE and TARGET_ASM_FUNCTION_EPILOGUE are
1636 The THUNK_FNDECL is redundant. (DELTA and FUNCTION have already been
1637 extracted from it.) It might possibly be useful on some targets, but
1640 If you do not define this macro, the target-independent code in the C++
1641 frontend will generate a less efficient heavyweight thunk that calls
1642 FUNCTION instead of jumping to it. The generic approach does not support
1644 #define ASM_OUTPUT_MI_THUNK(FILE, THUNK_FNDECL, DELTA, FUNCTION) \
1645 xstormy16_asm_output_mi_thunk (FILE, THUNK_FNDECL, DELTA, FUNCTION)
1648 /* Generating Code for Profiling. */
1650 /* A C statement or compound statement to output to FILE some assembler code to
1651 call the profiling subroutine `mcount'. Before calling, the assembler code
1652 must load the address of a counter variable into a register where `mcount'
1653 expects to find the address. The name of this variable is `LP' followed by
1654 the number LABELNO, so you would generate the name using `LP%d' in a
1657 The details of how the address should be passed to `mcount' are determined
1658 by your operating system environment, not by GNU CC. To figure them out,
1659 compile a small program for profiling using the system's installed C
1660 compiler and look at the assembler code that results.
1662 This declaration must be present, but it can be an abort if profiling is
1665 #define FUNCTION_PROFILER(FILE, LABELNO) abort ()
1667 /* Define this macro if the code for function profiling should come before the
1668 function prologue. Normally, the profiling code comes after. */
1669 /* #define PROFILE_BEFORE_PROLOGUE */
1672 /* If the target has particular reasons why a function cannot be inlined,
1673 it may define the TARGET_CANNOT_INLINE_P. This macro takes one argument,
1674 the DECL describing the function. The function should NULL if the function
1675 *can* be inlined. Otherwise it should return a pointer to a string containing
1676 a message describing why the function could not be inlined. The message will
1677 displayed if the '-Winline' command line switch has been given. If the message
1678 contains a '%s' sequence, this will be replaced by the name of the function. */
1679 /* #define TARGET_CANNOT_INLINE_P(FN_DECL) xstormy16_cannot_inline_p (FN_DECL) */
1681 /* Implementing the Varargs Macros. */
1683 /* If defined, is a C expression that produces the machine-specific code for a
1684 call to `__builtin_saveregs'. This code will be moved to the very beginning
1685 of the function, before any parameter access are made. The return value of
1686 this function should be an RTX that contains the value to use as the return
1687 of `__builtin_saveregs'.
1689 If this macro is not defined, the compiler will output an ordinary call to
1690 the library function `__builtin_saveregs'. */
1691 /* #define EXPAND_BUILTIN_SAVEREGS() */
1693 /* This macro offers an alternative to using `__builtin_saveregs' and defining
1694 the macro `EXPAND_BUILTIN_SAVEREGS'. Use it to store the anonymous register
1695 arguments into the stack so that all the arguments appear to have been
1696 passed consecutively on the stack. Once this is done, you can use the
1697 standard implementation of varargs that works for machines that pass all
1698 their arguments on the stack.
1700 The argument ARGS_SO_FAR is the `CUMULATIVE_ARGS' data structure, containing
1701 the values that obtain after processing of the named arguments. The
1702 arguments MODE and TYPE describe the last named argument--its machine mode
1703 and its data type as a tree node.
1705 The macro implementation should do two things: first, push onto the stack
1706 all the argument registers *not* used for the named arguments, and second,
1707 store the size of the data thus pushed into the `int'-valued variable whose
1708 name is supplied as the argument PRETEND_ARGS_SIZE. The value that you
1709 store here will serve as additional offset for setting up the stack frame.
1711 Because you must generate code to push the anonymous arguments at compile
1712 time without knowing their data types, `SETUP_INCOMING_VARARGS' is only
1713 useful on machines that have just a single category of argument register and
1714 use it uniformly for all data types.
1716 If the argument SECOND_TIME is nonzero, it means that the arguments of the
1717 function are being analyzed for the second time. This happens for an inline
1718 function, which is not actually compiled until the end of the source file.
1719 The macro `SETUP_INCOMING_VARARGS' should not generate any instructions in
1721 #define SETUP_INCOMING_VARARGS(ARGS_SO_FAR, MODE, TYPE, PRETEND_ARGS_SIZE, SECOND_TIME) \
1722 if (! SECOND_TIME) \
1723 xstormy16_setup_incoming_varargs (ARGS_SO_FAR, MODE, TYPE, & PRETEND_ARGS_SIZE)
1725 /* Define this macro if the location where a function argument is passed
1726 depends on whether or not it is a named argument.
1728 This macro controls how the NAMED argument to `FUNCTION_ARG' is set for
1729 varargs and stdarg functions. With this macro defined, the NAMED argument
1730 is always true for named arguments, and false for unnamed arguments. If
1731 this is not defined, but `SETUP_INCOMING_VARARGS' is defined, then all
1732 arguments are treated as named. Otherwise, all named arguments except the
1733 last are treated as named. */
1734 /* #define STRICT_ARGUMENT_NAMING 1 */
1736 /* Build up the stdarg/varargs va_list type tree, assinging it to NODE. If not
1737 defined, it is assumed that va_list is a void * pointer. */
1738 #define BUILD_VA_LIST_TYPE(NODE) \
1739 ((NODE) = xstormy16_build_va_list ())
1741 /* Implement the stdarg/varargs va_start macro. STDARG_P is non-zero if this
1742 is stdarg.h instead of varargs.h. VALIST is the tree of the va_list
1743 variable to initialize. NEXTARG is the machine independent notion of the
1744 'next' argument after the variable arguments. If not defined, a standard
1745 implementation will be defined that works for arguments passed on the stack. */
1746 #define EXPAND_BUILTIN_VA_START(STDARG_P, VALIST, NEXTARG) \
1747 xstormy16_expand_builtin_va_start (STDARG_P, VALIST, NEXTARG)
1749 /* Implement the stdarg/varargs va_arg macro. VALIST is the variable of type
1750 va_list as a tree, TYPE is the type passed to va_arg. */
1751 #define EXPAND_BUILTIN_VA_ARG(VALIST, TYPE) \
1752 xstormy16_expand_builtin_va_arg (VALIST, TYPE)
1754 /* Implement the stdarg/varargs va_end macro. VALIST is the variable of type
1755 va_list as a tree. */
1756 /* #define EXPAND_BUILTIN_VA_END(VALIST) */
1759 /* Trampolines for Nested Functions. */
1761 /* A C statement to output, on the stream FILE, assembler code for a block of
1762 data that contains the constant parts of a trampoline. This code should not
1763 include a label--the label is taken care of automatically. */
1764 /* #define TRAMPOLINE_TEMPLATE(FILE) */
1766 /* The name of a subroutine to switch to the section in which the trampoline
1767 template is to be placed. The default is a value of `readonly_data_section',
1768 which places the trampoline in the section containing read-only data. */
1769 /* #define TRAMPOLINE_SECTION */
1771 /* A C expression for the size in bytes of the trampoline, as an integer. */
1772 #define TRAMPOLINE_SIZE 8
1774 /* Alignment required for trampolines, in bits.
1776 If you don't define this macro, the value of `BIGGEST_ALIGNMENT' is used for
1777 aligning trampolines. */
1778 #define TRAMPOLINE_ALIGNMENT 16
1780 /* A C statement to initialize the variable parts of a trampoline. ADDR is an
1781 RTX for the address of the trampoline; FNADDR is an RTX for the address of
1782 the nested function; STATIC_CHAIN is an RTX for the static chain value that
1783 should be passed to the function when it is called. */
1784 #define INITIALIZE_TRAMPOLINE(ADDR, FNADDR, STATIC_CHAIN) \
1785 xstormy16_initialize_trampoline (ADDR, FNADDR, STATIC_CHAIN)
1787 /* A C expression to allocate run-time space for a trampoline. The expression
1788 value should be an RTX representing a memory reference to the space for the
1791 If this macro is not defined, by default the trampoline is allocated as a
1792 stack slot. This default is right for most machines. The exceptions are
1793 machines where it is impossible to execute instructions in the stack area.
1794 On such machines, you may have to implement a separate stack, using this
1795 macro in conjunction with `TARGET_ASM_FUNCTION_PROLOGUE' and
1796 `TARGET_ASM_FUNCTION_EPILOGUE'.
1798 FP points to a data structure, a `struct function', which describes the
1799 compilation status of the immediate containing function of the function
1800 which the trampoline is for. Normally (when `ALLOCATE_TRAMPOLINE' is not
1801 defined), the stack slot for the trampoline is in the stack frame of this
1802 containing function. Other allocation strategies probably must do something
1803 analogous with this information. */
1804 /* #define ALLOCATE_TRAMPOLINE(FP) */
1806 /* Implementing trampolines is difficult on many machines because they have
1807 separate instruction and data caches. Writing into a stack location fails
1808 to clear the memory in the instruction cache, so when the program jumps to
1809 that location, it executes the old contents.
1811 Here are two possible solutions. One is to clear the relevant parts of the
1812 instruction cache whenever a trampoline is set up. The other is to make all
1813 trampolines identical, by having them jump to a standard subroutine. The
1814 former technique makes trampoline execution faster; the latter makes
1815 initialization faster.
1817 To clear the instruction cache when a trampoline is initialized, define the
1818 following macros which describe the shape of the cache. */
1820 /* The total size in bytes of the cache. */
1821 /* #define INSN_CACHE_SIZE */
1823 /* The length in bytes of each cache line. The cache is divided into cache
1824 lines which are disjoint slots, each holding a contiguous chunk of data
1825 fetched from memory. Each time data is brought into the cache, an entire
1826 line is read at once. The data loaded into a cache line is always aligned
1827 on a boundary equal to the line size. */
1828 /* #define INSN_CACHE_LINE_WIDTH */
1830 /* The number of alternative cache lines that can hold any particular memory
1832 /* #define INSN_CACHE_DEPTH */
1834 /* Alternatively, if the machine has system calls or instructions to clear the
1835 instruction cache directly, you can define the following macro. */
1837 /* If defined, expands to a C expression clearing the *instruction cache* in
1838 the specified interval. If it is not defined, and the macro INSN_CACHE_SIZE
1839 is defined, some generic code is generated to clear the cache. The
1840 definition of this macro would typically be a series of `asm' statements.
1841 Both BEG and END are both pointer expressions. */
1842 /* #define CLEAR_INSN_CACHE (BEG, END) */
1844 /* To use a standard subroutine, define the following macro. In addition, you
1845 must make sure that the instructions in a trampoline fill an entire cache
1846 line with identical instructions, or else ensure that the beginning of the
1847 trampoline code is always aligned at the same point in its cache line. Look
1848 in `m68k.h' as a guide. */
1850 /* Define this macro if trampolines need a special subroutine to do their work.
1851 The macro should expand to a series of `asm' statements which will be
1852 compiled with GNU CC. They go in a library function named
1853 `__transfer_from_trampoline'.
1855 If you need to avoid executing the ordinary prologue code of a compiled C
1856 function when you jump to the subroutine, you can do so by placing a special
1857 label of your own in the assembler code. Use one `asm' statement to
1858 generate an assembler label, and another to make the label global. Then
1859 trampolines can use that label to jump directly to your special assembler
1861 /* #define TRANSFER_FROM_TRAMPOLINE */
1864 /* Implicit Calls to Library Routines */
1866 /* A C string constant giving the name of the function to call for
1867 multiplication of one signed full-word by another. If you do not define
1868 this macro, the default name is used, which is `__mulsi3', a function
1869 defined in `libgcc.a'. */
1870 /* #define MULSI3_LIBCALL */
1872 /* A C string constant giving the name of the function to call for division of
1873 one signed full-word by another. If you do not define this macro, the
1874 default name is used, which is `__divsi3', a function defined in `libgcc.a'. */
1875 /* #define DIVSI3_LIBCALL */
1877 /* A C string constant giving the name of the function to call for division of
1878 one unsigned full-word by another. If you do not define this macro, the
1879 default name is used, which is `__udivsi3', a function defined in
1881 /* #define UDIVSI3_LIBCALL */
1883 /* A C string constant giving the name of the function to call for the
1884 remainder in division of one signed full-word by another. If you do not
1885 define this macro, the default name is used, which is `__modsi3', a function
1886 defined in `libgcc.a'. */
1887 /* #define MODSI3_LIBCALL */
1889 /* A C string constant giving the name of the function to call for the
1890 remainder in division of one unsigned full-word by another. If you do not
1891 define this macro, the default name is used, which is `__umodsi3', a
1892 function defined in `libgcc.a'. */
1893 /* #define UMODSI3_LIBCALL */
1895 /* A C string constant giving the name of the function to call for
1896 multiplication of one signed double-word by another. If you do not define
1897 this macro, the default name is used, which is `__muldi3', a function
1898 defined in `libgcc.a'. */
1899 /* #define MULDI3_LIBCALL */
1901 /* A C string constant giving the name of the function to call for division of
1902 one signed double-word by another. If you do not define this macro, the
1903 default name is used, which is `__divdi3', a function defined in `libgcc.a'. */
1904 /* #define DIVDI3_LIBCALL */
1906 /* A C string constant giving the name of the function to call for division of
1907 one unsigned full-word by another. If you do not define this macro, the
1908 default name is used, which is `__udivdi3', a function defined in
1910 /* #define UDIVDI3_LIBCALL */
1912 /* A C string constant giving the name of the function to call for the
1913 remainder in division of one signed double-word by another. If you do not
1914 define this macro, the default name is used, which is `__moddi3', a function
1915 defined in `libgcc.a'. */
1916 /* #define MODDI3_LIBCALL */
1918 /* A C string constant giving the name of the function to call for the
1919 remainder in division of one unsigned full-word by another. If you do not
1920 define this macro, the default name is used, which is `__umoddi3', a
1921 function defined in `libgcc.a'. */
1922 /* #define UMODDI3_LIBCALL */
1924 /* Define this macro as a C statement that declares additional library routines
1925 renames existing ones. `init_optabs' calls this macro after initializing all
1926 the normal library routines. */
1927 /* #define INIT_TARGET_OPTABS */
1929 /* The value of `EDOM' on the target machine, as a C integer constant
1930 expression. If you don't define this macro, GNU CC does not attempt to
1931 deposit the value of `EDOM' into `errno' directly. Look in
1932 `/usr/include/errno.h' to find the value of `EDOM' on your system.
1934 If you do not define `TARGET_EDOM', then compiled code reports domain errors
1935 by calling the library function and letting it report the error. If
1936 mathematical functions on your system use `matherr' when there is an error,
1937 then you should leave `TARGET_EDOM' undefined so that `matherr' is used
1939 /* #define TARGET_EDOM */
1941 /* Define this macro as a C expression to create an rtl expression that refers
1942 to the global "variable" `errno'. (On certain systems, `errno' may not
1943 actually be a variable.) If you don't define this macro, a reasonable
1945 /* #define GEN_ERRNO_RTX */
1947 /* Define this macro if GNU CC should generate calls to the System V (and ANSI
1948 C) library functions `memcpy' and `memset' rather than the BSD functions
1949 `bcopy' and `bzero'.
1951 Defined in svr4.h. */
1952 #define TARGET_MEM_FUNCTIONS
1954 /* Define this macro if only `float' arguments cannot be passed to library
1955 routines (so they must be converted to `double'). This macro affects both
1956 how library calls are generated and how the library routines in `libgcc1.c'
1957 accept their arguments. It is useful on machines where floating and fixed
1958 point arguments are passed differently, such as the i860. */
1959 /* #define LIBGCC_NEEDS_DOUBLE */
1961 /* Define this macro to override the type used by the library routines to pick
1962 up arguments of type `float'. (By default, they use a union of `float' and
1965 The obvious choice would be `float'--but that won't work with traditional C
1966 compilers that expect all arguments declared as `float' to arrive as
1967 `double'. To avoid this conversion, the library routines ask for the value
1968 as some other type and then treat it as a `float'.
1970 On some systems, no other type will work for this. For these systems, you
1971 must use `LIBGCC_NEEDS_DOUBLE' instead, to force conversion of the values
1972 `double' before they are passed. */
1973 /* #define FLOAT_ARG_TYPE */
1975 /* Define this macro to override the way library routines redesignate a `float'
1976 argument as a `float' instead of the type it was passed as. The default is
1977 an expression which takes the `float' field of the union. */
1978 /* #define FLOATIFY(PASSED_VALUE) */
1980 /* Define this macro to override the type used by the library routines to
1981 return values that ought to have type `float'. (By default, they use
1984 The obvious choice would be `float'--but that won't work with traditional C
1985 compilers gratuitously convert values declared as `float' into `double'. */
1986 /* #define FLOAT_VALUE_TYPE */
1988 /* Define this macro to override the way the value of a `float'-returning
1989 library routine should be packaged in order to return it. These functions
1990 are actually declared to return type `FLOAT_VALUE_TYPE' (normally `int').
1992 These values can't be returned as type `float' because traditional C
1993 compilers would gratuitously convert the value to a `double'.
1995 A local variable named `intify' is always available when the macro `INTIFY'
1996 is used. It is a union of a `float' field named `f' and a field named `i'
1997 whose type is `FLOAT_VALUE_TYPE' or `int'.
1999 If you don't define this macro, the default definition works by copying the
2000 value through that union. */
2001 /* #define INTIFY(FLOAT_VALUE) */
2003 /* Define this macro as the name of the data type corresponding to `SImode' in
2004 the system's own C compiler.
2006 You need not define this macro if that type is `long int', as it usually is. */
2007 /* #define nongcc_SI_type */
2009 /* Define this macro as the name of the data type corresponding to the
2010 word_mode in the system's own C compiler.
2012 You need not define this macro if that type is `long int', as it usually is. */
2013 /* #define nongcc_word_type */
2015 /* Define these macros to supply explicit C statements to carry out various
2016 arithmetic operations on types `float' and `double' in the library routines
2017 in `libgcc1.c'. See that file for a full list of these macros and their
2020 On most machines, you don't need to define any of these macros, because the
2021 C compiler that comes with the system takes care of doing them. */
2022 /* #define perform_... */
2024 /* Define this macro to generate code for Objective C message sending using the
2025 calling convention of the NeXT system. This calling convention involves
2026 passing the object, the selector and the method arguments all at once to the
2027 method-lookup library function.
2029 The default calling convention passes just the object and the selector to
2030 the lookup function, which returns a pointer to the method. */
2031 /* #define NEXT_OBJC_RUNTIME */
2034 /* Addressing Modes */
2036 /* Define this macro if the machine supports post-increment addressing. */
2037 #define HAVE_POST_INCREMENT 1
2039 /* Similar for other kinds of addressing. */
2040 /* #define HAVE_PRE_INCREMENT 1 */
2041 /* #define HAVE_POST_DECREMENT 1 */
2042 #define HAVE_PRE_DECREMENT 1
2044 /* A C expression that is 1 if the RTX X is a constant which is a valid
2045 address. On most machines, this can be defined as `CONSTANT_P (X)', but a
2046 few machines are more restrictive in which constant addresses are supported.
2048 `CONSTANT_P' accepts integer-values expressions whose values are not
2049 explicitly known, such as `symbol_ref', `label_ref', and `high' expressions
2050 and `const' arithmetic expressions, in addition to `const_int' and
2051 `const_double' expressions. */
2052 #define CONSTANT_ADDRESS_P(X) CONSTANT_P (X)
2054 /* A number, the maximum number of registers that can appear in a valid memory
2055 address. Note that it is up to you to specify a value equal to the maximum
2056 number that `GO_IF_LEGITIMATE_ADDRESS' would ever accept. */
2057 #define MAX_REGS_PER_ADDRESS 1
2059 /* A C compound statement with a conditional `goto LABEL;' executed if X (an
2060 RTX) is a legitimate memory address on the target machine for a memory
2061 operand of mode MODE. */
2062 #ifdef REG_OK_STRICT
2063 #define GO_IF_LEGITIMATE_ADDRESS(MODE, X, LABEL) \
2065 if (xstormy16_legitimate_address_p (MODE, X, 1)) \
2069 #define GO_IF_LEGITIMATE_ADDRESS(MODE, X, LABEL) \
2071 if (xstormy16_legitimate_address_p (MODE, X, 0)) \
2075 /* A C expression that is nonzero if X (assumed to be a `reg' RTX) is valid for
2076 use as a base register. For hard registers, it should always accept those
2077 which the hardware permits and reject the others. Whether the macro accepts
2078 or rejects pseudo registers must be controlled by `REG_OK_STRICT' as
2079 described above. This usually requires two variant definitions, of which
2080 `REG_OK_STRICT' controls the one actually used. */
2081 #ifdef REG_OK_STRICT
2082 #define REG_OK_FOR_BASE_P(X) \
2083 (REGNO_OK_FOR_BASE_P (REGNO (X)) && (REGNO (X) < FIRST_PSEUDO_REGISTER))
2085 #define REG_OK_FOR_BASE_P(X) REGNO_OK_FOR_BASE_P (REGNO (X))
2088 /* A C expression that is nonzero if X (assumed to be a `reg' RTX) is valid for
2089 use as an index register.
2091 The difference between an index register and a base register is that the
2092 index register may be scaled. If an address involves the sum of two
2093 registers, neither one of them scaled, then either one may be labeled the
2094 "base" and the other the "index"; but whichever labeling is used must fit
2095 the machine's constraints of which registers may serve in each capacity.
2096 The compiler will try both labelings, looking for one that is valid, and
2097 will reload one or both registers only if neither labeling works. */
2098 #define REG_OK_FOR_INDEX_P(X) REG_OK_FOR_BASE_P (X)
2100 /* A C compound statement that attempts to replace X with a valid memory
2101 address for an operand of mode MODE. WIN will be a C statement label
2102 elsewhere in the code; the macro definition may use
2104 GO_IF_LEGITIMATE_ADDRESS (MODE, X, WIN);
2106 to avoid further processing if the address has become legitimate.
2108 X will always be the result of a call to `break_out_memory_refs', and OLDX
2109 will be the operand that was given to that function to produce X.
2111 The code generated by this macro should not alter the substructure of X. If
2112 it transforms X into a more legitimate form, it should assign X (which will
2113 always be a C variable) a new value.
2115 It is not necessary for this macro to come up with a legitimate address.
2116 The compiler has standard ways of doing so in all cases. In fact, it is
2117 safe for this macro to do nothing. But often a machine-dependent strategy
2118 can generate better code. */
2119 #define LEGITIMIZE_ADDRESS(X, OLDX, MODE, WIN)
2121 /* A C statement or compound statement with a conditional `goto LABEL;'
2122 executed if memory address X (an RTX) can have different meanings depending
2123 on the machine mode of the memory reference it is used for or if the address
2124 is valid for some modes but not others.
2126 Autoincrement and autodecrement addresses typically have mode-dependent
2127 effects because the amount of the increment or decrement is the size of the
2128 operand being addressed. Some machines have other mode-dependent addresses.
2129 Many RISC machines have no mode-dependent addresses.
2131 You may assume that ADDR is a valid address for the machine.
2133 On this chip, this is true if the address is valid with an offset
2134 of 0 but not of 6, because in that case it cannot be used as an
2135 address for DImode or DFmode, or if the address is a post-increment
2136 or pre-decrement address.
2138 #define GO_IF_MODE_DEPENDENT_ADDRESS(ADDR,LABEL) \
2139 if (xstormy16_mode_dependent_address_p (ADDR)) \
2142 /* A C expression that is nonzero if X is a legitimate constant for an
2143 immediate operand on the target machine. You can assume that X satisfies
2144 `CONSTANT_P', so you need not check this. In fact, `1' is a suitable
2145 definition for this macro on machines where anything `CONSTANT_P' is valid. */
2146 #define LEGITIMATE_CONSTANT_P(X) 1
2149 /* Condition Code Status */
2151 /* C code for a data type which is used for declaring the `mdep' component of
2152 `cc_status'. It defaults to `int'.
2154 This macro is not used on machines that do not use `cc0'. */
2155 /* #define CC_STATUS_MDEP */
2157 /* A C expression to initialize the `mdep' field to "empty". The default
2158 definition does nothing, since most machines don't use the field anyway. If
2159 you want to use the field, you should probably define this macro to
2162 This macro is not used on machines that do not use `cc0'. */
2163 /* #define CC_STATUS_MDEP_INIT */
2165 /* A C compound statement to set the components of `cc_status' appropriately
2166 for an insn INSN whose body is EXP. It is this macro's responsibility to
2167 recognize insns that set the condition code as a byproduct of other activity
2168 as well as those that explicitly set `(cc0)'.
2170 This macro is not used on machines that do not use `cc0'.
2172 If there are insns that do not set the condition code but do alter other
2173 machine registers, this macro must check to see whether they invalidate the
2174 expressions that the condition code is recorded as reflecting. For example,
2175 on the 68000, insns that store in address registers do not set the condition
2176 code, which means that usually `NOTICE_UPDATE_CC' can leave `cc_status'
2177 unaltered for such insns. But suppose that the previous insn set the
2178 condition code based on location `a4@(102)' and the current insn stores a
2179 new value in `a4'. Although the condition code is not changed by this, it
2180 will no longer be true that it reflects the contents of `a4@(102)'.
2181 Therefore, `NOTICE_UPDATE_CC' must alter `cc_status' in this case to say
2182 that nothing is known about the condition code value.
2184 The definition of `NOTICE_UPDATE_CC' must be prepared to deal with the
2185 results of peephole optimization: insns whose patterns are `parallel' RTXs
2186 containing various `reg', `mem' or constants which are just the operands.
2187 The RTL structure of these insns is not sufficient to indicate what the
2188 insns actually do. What `NOTICE_UPDATE_CC' should do when it sees one is
2189 just to run `CC_STATUS_INIT'.
2191 A possible definition of `NOTICE_UPDATE_CC' is to call a function that looks
2192 at an attribute named, for example, `cc'. This avoids having detailed
2193 information about patterns in two places, the `md' file and in
2194 `NOTICE_UPDATE_CC'. */
2195 /* #define NOTICE_UPDATE_CC(EXP, INSN) */
2197 /* A list of names to be used for additional modes for condition code values in
2198 registers. These names are added to `enum machine_mode' and all have class
2199 `MODE_CC'. By convention, they should start with `CC' and end with `mode'.
2201 You should only define this macro if your machine does not use `cc0' and
2202 only if additional modes are required. */
2203 /* #define EXTRA_CC_MODES */
2205 /* Returns a mode from class `MODE_CC' to be used when comparison operation
2206 code OP is applied to rtx X and Y. For example, on the Sparc,
2207 `SELECT_CC_MODE' is defined as (see *note Jump Patterns::. for a
2208 description of the reason for this definition)
2210 #define SELECT_CC_MODE(OP,X,Y) \
2211 (GET_MODE_CLASS (GET_MODE (X)) == MODE_FLOAT \
2212 ? ((OP == EQ || OP == NE) ? CCFPmode : CCFPEmode) \
2213 : ((GET_CODE (X) == PLUS || GET_CODE (X) == MINUS \
2214 || GET_CODE (X) == NEG) \
2215 ? CC_NOOVmode : CCmode))
2217 You need not define this macro if `EXTRA_CC_MODES' is not defined. */
2218 /* #define SELECT_CC_MODE(OP, X, Y) */
2220 /* One some machines not all possible comparisons are defined, but you can
2221 convert an invalid comparison into a valid one. For example, the Alpha does
2222 not have a `GT' comparison, but you can use an `LT' comparison instead and
2223 swap the order of the operands.
2225 On such machines, define this macro to be a C statement to do any required
2226 conversions. CODE is the initial comparison code and OP0 and OP1 are the
2227 left and right operands of the comparison, respectively. You should modify
2228 CODE, OP0, and OP1 as required.
2230 GNU CC will not assume that the comparison resulting from this macro is
2231 valid but will see if the resulting insn matches a pattern in the `md' file.
2233 You need not define this macro if it would never change the comparison code
2235 /* #define CANONICALIZE_COMPARISON(CODE, OP0, OP1) */
2237 /* A C expression whose value is one if it is always safe to reverse a
2238 comparison whose mode is MODE. If `SELECT_CC_MODE' can ever return MODE for
2239 a floating-point inequality comparison, then `REVERSIBLE_CC_MODE (MODE)'
2242 You need not define this macro if it would always returns zero or if the
2243 floating-point format is anything other than `IEEE_FLOAT_FORMAT'. For
2244 example, here is the definition used on the Sparc, where floating-point
2245 inequality comparisons are always given `CCFPEmode':
2247 #define REVERSIBLE_CC_MODE(MODE) ((MODE) != CCFPEmode) */
2248 /* #define REVERSIBLE_CC_MODE(MODE) */
2251 /* Describing Relative Costs of Operations */
2253 /* A part of a C `switch' statement that describes the relative costs of
2254 constant RTL expressions. It must contain `case' labels for expression
2255 codes `const_int', `const', `symbol_ref', `label_ref' and `const_double'.
2256 Each case must ultimately reach a `return' statement to return the relative
2257 cost of the use of that kind of constant value in an expression. The cost
2258 may depend on the precise value of the constant, which is available for
2259 examination in X, and the rtx code of the expression in which it is
2260 contained, found in OUTER_CODE.
2262 CODE is the expression code--redundant, since it can be obtained with
2264 #define CONST_COSTS(X, CODE, OUTER_CODE) \
2266 if (INTVAL (X) < 16 && INTVAL (X) >= 0) \
2267 return COSTS_N_INSNS (1)/2; \
2268 if (INTVAL (X) < 256 && INTVAL (X) >= 0) \
2269 return COSTS_N_INSNS (1); \
2270 case CONST_DOUBLE: \
2274 return COSTS_N_INSNS(2);
2276 /* Like `CONST_COSTS' but applies to nonconstant RTL expressions. This can be
2277 used, for example, to indicate how costly a multiply instruction is. In
2278 writing this macro, you can use the construct `COSTS_N_INSNS (N)' to specify
2279 a cost equal to N fast instructions. OUTER_CODE is the code of the
2280 expression in which X is contained.
2282 This macro is optional; do not define it if the default cost assumptions are
2283 adequate for the target machine. */
2284 #define RTX_COSTS(X, CODE, OUTER_CODE) \
2286 return COSTS_N_INSNS (35 + 6); \
2288 return COSTS_N_INSNS (51 - 6);
2290 /* An expression giving the cost of an addressing mode that contains ADDRESS.
2291 If not defined, the cost is computed from the ADDRESS expression and the
2292 `CONST_COSTS' values.
2294 For most CISC machines, the default cost is a good approximation of the true
2295 cost of the addressing mode. However, on RISC machines, all instructions
2296 normally have the same length and execution time. Hence all addresses will
2299 In cases where more than one form of an address is known, the form with the
2300 lowest cost will be used. If multiple forms have the same, lowest, cost,
2301 the one that is the most complex will be used.
2303 For example, suppose an address that is equal to the sum of a register and a
2304 constant is used twice in the same basic block. When this macro is not
2305 defined, the address will be computed in a register and memory references
2306 will be indirect through that register. On machines where the cost of the
2307 addressing mode containing the sum is no higher than that of a simple
2308 indirect reference, this will produce an additional instruction and possibly
2309 require an additional register. Proper specification of this macro
2310 eliminates this overhead for such machines.
2312 Similar use of this macro is made in strength reduction of loops.
2314 ADDRESS need not be valid as an address. In such a case, the cost is not
2315 relevant and can be any value; invalid addresses need not be assigned a
2318 On machines where an address involving more than one register is as cheap as
2319 an address computation involving only one register, defining `ADDRESS_COST'
2320 to reflect this can cause two registers to be live over a region of code
2321 where only one would have been if `ADDRESS_COST' were not defined in that
2322 manner. This effect should be considered in the definition of this macro.
2323 Equivalent costs should probably only be given to addresses with different
2324 numbers of registers on machines with lots of registers.
2326 This macro will normally either not be defined or be defined as a
2328 #define ADDRESS_COST(ADDRESS) \
2329 (GET_CODE (ADDRESS) == CONST_INT ? 2 \
2330 : GET_CODE (ADDRESS) == PLUS ? 7 \
2333 /* A C expression for the cost of moving data of mode MODE from a
2334 register in class FROM to one in class TO. The classes are
2335 expressed using the enumeration values such as `GENERAL_REGS'. A
2336 value of 4 is the default; other values are interpreted relative to
2339 It is not required that the cost always equal 2 when FROM is the same as TO;
2340 on some machines it is expensive to move between registers if they are not
2343 If reload sees an insn consisting of a single `set' between two hard
2344 registers, and if `REGISTER_MOVE_COST' applied to their classes returns a
2345 value of 2, reload does not check to ensure that the constraints of the insn
2346 are met. Setting a cost of other than 2 will allow reload to verify that
2347 the constraints are met. You should do this if the `movM' pattern's
2348 constraints do not allow such copying. */
2349 #define REGISTER_MOVE_COST(MODE, FROM, TO) 2
2351 /* A C expression for the cost of moving data of mode M between a register and
2352 memory. A value of 2 is the default; this cost is relative to those in
2353 `REGISTER_MOVE_COST'.
2355 If moving between registers and memory is more expensive than between two
2356 registers, you should define this macro to express the relative cost. */
2357 #define MEMORY_MOVE_COST(M,C,I) (5 + memory_move_secondary_cost (M, C, I))
2359 /* A C expression for the cost of a branch instruction. A value of 1 is the
2360 default; other values are interpreted relative to that. */
2362 #define BRANCH_COST 5
2364 /* Here are additional macros which do not specify precise relative costs, but
2365 only that certain actions are more expensive than GNU CC would ordinarily
2368 /* Define this macro as a C expression which is nonzero if accessing less than
2369 a word of memory (i.e. a `char' or a `short') is no faster than accessing a
2370 word of memory, i.e., if such access require more than one instruction or if
2371 there is no difference in cost between byte and (aligned) word loads.
2373 When this macro is not defined, the compiler will access a field by finding
2374 the smallest containing object; when it is defined, a fullword load will be
2375 used if alignment permits. Unless bytes accesses are faster than word
2376 accesses, using word accesses is preferable since it may eliminate
2377 subsequent memory access if subsequent accesses occur to other fields in the
2378 same word of the structure, but to different bytes. */
2379 #define SLOW_BYTE_ACCESS 0
2381 /* Define this macro to be the value 1 if unaligned accesses have a cost many
2382 times greater than aligned accesses, for example if they are emulated in a
2385 When this macro is non-zero, the compiler will act as if `STRICT_ALIGNMENT'
2386 were non-zero when generating code for block moves. This can cause
2387 significantly more instructions to be produced. Therefore, do not set this
2388 macro non-zero if unaligned accesses only add a cycle or two to the time for
2391 If the value of this macro is always zero, it need not be defined. */
2392 /* #define SLOW_UNALIGNED_ACCESS */
2394 /* Define this macro to inhibit strength reduction of memory addresses. (On
2395 some machines, such strength reduction seems to do harm rather than good.) */
2396 /* #define DONT_REDUCE_ADDR */
2398 /* The number of scalar move insns which should be generated instead of a
2399 string move insn or a library call. Increasing the value will always make
2400 code faster, but eventually incurs high cost in increased code size.
2402 If you don't define this, a reasonable default is used. */
2403 /* #define MOVE_RATIO */
2405 /* Define this macro if it is as good or better to call a constant function
2406 address than to call an address kept in a register. */
2407 #define NO_FUNCTION_CSE
2409 /* Define this macro if it is as good or better for a function to call itself
2410 with an explicit address than to call an address kept in a register. */
2411 #define NO_RECURSIVE_FUNCTION_CSE
2413 /* A C statement (sans semicolon) to update the integer variable COST based on
2414 the relationship between INSN that is dependent on DEP_INSN through the
2415 dependence LINK. The default is to make no adjustment to COST. This can be
2416 used for example to specify to the scheduler that an output- or
2417 anti-dependence does not incur the same cost as a data-dependence. */
2418 /* #define ADJUST_COST(INSN, LINK, DEP_INSN, COST) */
2420 /* A C statement (sans semicolon) to update the integer scheduling
2421 priority `INSN_PRIORITY(INSN)'. Reduce the priority to execute
2422 the INSN earlier, increase the priority to execute INSN later.
2423 Do not define this macro if you do not need to adjust the
2424 scheduling priorities of insns. */
2425 /* #define ADJUST_PRIORITY (INSN) */
2428 /* Dividing the output into sections. */
2430 /* A C expression whose value is a string containing the assembler operation
2431 that should precede instructions and read-only data. Normally `".text"' is
2433 #define TEXT_SECTION_ASM_OP ".text"
2435 /* A C expression whose value is a string containing the assembler operation to
2436 identify the following data as writable initialized data. Normally
2437 `".data"' is right. */
2438 #define DATA_SECTION_ASM_OP ".data"
2440 /* if defined, a C expression whose value is a string containing the assembler
2441 operation to identify the following data as shared data. If not defined,
2442 `DATA_SECTION_ASM_OP' will be used. */
2443 /* #define SHARED_SECTION_ASM_OP */
2445 /* If defined, a C expression whose value is a string containing the
2446 assembler operation to identify the following data as
2447 uninitialized global data. If not defined, and neither
2448 `ASM_OUTPUT_BSS' nor `ASM_OUTPUT_ALIGNED_BSS' are defined,
2449 uninitialized global data will be output in the data section if
2450 `-fno-common' is passed, otherwise `ASM_OUTPUT_COMMON' will be
2452 #define BSS_SECTION_ASM_OP ".bss"
2454 /* If defined, a C expression whose value is a string containing the
2455 assembler operation to identify the following data as
2456 uninitialized global shared data. If not defined, and
2457 `BSS_SECTION_ASM_OP' is, the latter will be used. */
2458 /* #define SHARED_BSS_SECTION_ASM_OP */
2460 /* Define the pseudo-ops used to switch to the .ctors and .dtors sections.
2461 There are no shared libraries on this target so these sections need
2464 Defined in elfos.h. */
2466 #undef CTORS_SECTION_ASM_OP
2467 #undef DTORS_SECTION_ASM_OP
2468 #define CTORS_SECTION_ASM_OP "\t.section\t.ctors,\"a\""
2469 #define DTORS_SECTION_ASM_OP "\t.section\t.dtors,\"a\""
2471 /* A list of names for sections other than the standard two, which are
2472 `in_text' and `in_data'. You need not define this macro on a system with no
2473 other sections (that GCC needs to use).
2475 Defined in svr4.h. */
2476 /* #define EXTRA_SECTIONS */
2478 /* One or more functions to be defined in `varasm.c'. These functions should
2479 do jobs analogous to those of `text_section' and `data_section', for your
2480 additional sections. Do not define this macro if you do not define
2483 Defined in svr4.h. */
2484 /* #define EXTRA_SECTION_FUNCTIONS */
2486 /* Define this macro if jump tables (for `tablejump' insns) should be output in
2487 the text section, along with the assembler instructions. Otherwise, the
2488 readonly data section is used.
2490 This macro is irrelevant if there is no separate readonly data section. */
2491 #define JUMP_TABLES_IN_TEXT_SECTION 1
2493 /* Position Independent Code. */
2495 /* The register number of the register used to address a table of static data
2496 addresses in memory. In some cases this register is defined by a
2497 processor's "application binary interface" (ABI). When this macro is
2498 defined, RTL is generated for this register once, as with the stack pointer
2499 and frame pointer registers. If this macro is not defined, it is up to the
2500 machine-dependent files to allocate such a register (if necessary). */
2501 /* #define PIC_OFFSET_TABLE_REGNUM */
2503 /* Define this macro if the register defined by `PIC_OFFSET_TABLE_REGNUM' is
2504 clobbered by calls. Do not define this macro if `PPIC_OFFSET_TABLE_REGNUM'
2506 /* #define PIC_OFFSET_TABLE_REG_CALL_CLOBBERED */
2508 /* By generating position-independent code, when two different programs (A and
2509 B) share a common library (libC.a), the text of the library can be shared
2510 whether or not the library is linked at the same address for both programs.
2511 In some of these environments, position-independent code requires not only
2512 the use of different addressing modes, but also special code to enable the
2513 use of these addressing modes.
2515 The `FINALIZE_PIC' macro serves as a hook to emit these special codes once
2516 the function is being compiled into assembly code, but not before. (It is
2517 not done before, because in the case of compiling an inline function, it
2518 would lead to multiple PIC prologues being included in functions which used
2519 inline functions and were compiled to assembly language.) */
2520 /* #define FINALIZE_PIC */
2522 /* A C expression that is nonzero if X is a legitimate immediate operand on the
2523 target machine when generating position independent code. You can assume
2524 that X satisfies `CONSTANT_P', so you need not check this. You can also
2525 assume FLAG_PIC is true, so you need not check it either. You need not
2526 define this macro if all constants (including `SYMBOL_REF') can be immediate
2527 operands when generating position independent code. */
2528 /* #define LEGITIMATE_PIC_OPERAND_P(X) */
2531 /* The Overall Framework of an Assembler File. */
2533 /* A C expression which outputs to the stdio stream STREAM some appropriate
2534 text to go at the start of an assembler file.
2536 Normally this macro is defined to output a line containing `#NO_APP', which
2537 is a comment that has no effect on most assemblers but tells the GNU
2538 assembler that it can save time by not checking for certain assembler
2541 On systems that use SDB, it is necessary to output certain commands; see
2544 Defined in svr4.h. */
2545 /* #define ASM_FILE_START(STREAM) */
2547 /* A C expression which outputs to the stdio stream STREAM some appropriate
2548 text to go at the end of an assembler file.
2550 If this macro is not defined, the default is to output nothing special at
2551 the end of the file. Most systems don't require any definition.
2553 On systems that use SDB, it is necessary to output certain commands; see
2556 Defined in svr4.h. */
2557 /* #define ASM_FILE_END(STREAM) */
2559 /* A C string constant describing how to begin a comment in the target
2560 assembler language. The compiler assumes that the comment will end at the
2562 #define ASM_COMMENT_START ";"
2564 /* A C string constant for text to be output before each `asm' statement or
2565 group of consecutive ones. Normally this is `"#APP"', which is a comment
2566 that has no effect on most assemblers but tells the GNU assembler that it
2567 must check the lines that follow for all valid assembler constructs. */
2568 #define ASM_APP_ON "#APP\n"
2570 /* A C string constant for text to be output after each `asm' statement or
2571 group of consecutive ones. Normally this is `"#NO_APP"', which tells the
2572 GNU assembler to resume making the time-saving assumptions that are valid
2573 for ordinary compiler output. */
2574 #define ASM_APP_OFF "#NO_APP\n"
2576 /* A C statement to output COFF information or DWARF debugging information
2577 which indicates that filename NAME is the current source file to the stdio
2580 This macro need not be defined if the standard form of output for the file
2581 format in use is appropriate. */
2582 /* #define ASM_OUTPUT_SOURCE_FILENAME(STREAM, NAME) */
2584 /* A C statement to output DBX or SDB debugging information before code for
2585 line number LINE of the current source file to the stdio stream STREAM.
2587 This macro need not be defined if the standard form of debugging information
2588 for the debugger in use is appropriate.
2590 Defined in svr4.h. */
2591 /* #define ASM_OUTPUT_SOURCE_LINE(STREAM, LINE) */
2593 /* A C statement to output something to the assembler file to handle a `#ident'
2594 directive containing the text STRING. If this macro is not defined, nothing
2595 is output for a `#ident' directive.
2597 Defined in svr4.h. */
2598 /* #define ASM_OUTPUT_IDENT(STREAM, STRING) */
2600 /* A C statement to output something to the assembler file to switch to section
2601 NAME for object DECL which is either a `FUNCTION_DECL', a `VAR_DECL' or
2602 `NULL_TREE'. Some target formats do not support arbitrary sections. Do not
2603 define this macro in such cases.
2605 At present this macro is only used to support section attributes. When this
2606 macro is undefined, section attributes are disabled.
2608 Defined in svr4.h. */
2609 /* #define ASM_OUTPUT_SECTION_NAME(STREAM, DECL, NAME) */
2611 /* A C statement to output any assembler statements which are required to
2612 precede any Objective C object definitions or message sending. The
2613 statement is executed only when compiling an Objective C program. */
2614 /* #define OBJC_PROLOGUE */
2617 /* Output of Data. */
2619 /* A C statement to output to the stdio stream STREAM an assembler instruction
2620 to assemble a string constant containing the LEN bytes at PTR. PTR will be
2621 a C expression of type `char *' and LEN a C expression of type `int'.
2623 If the assembler has a `.ascii' pseudo-op as found in the Berkeley Unix
2624 assembler, do not define the macro `ASM_OUTPUT_ASCII'.
2626 Defined in svr4.h. */
2627 /* #define ASM_OUTPUT_ASCII(STREAM, PTR, LEN) */
2629 /* You may define this macro as a C expression. You should define the
2630 expression to have a non-zero value if GNU CC should output the
2631 constant pool for a function before the code for the function, or
2632 a zero value if GNU CC should output the constant pool after the
2633 function. If you do not define this macro, the usual case, GNU CC
2634 will output the constant pool before the function. */
2635 /* #define CONSTANT_POOL_BEFORE_FUNCTION */
2637 /* A C statement to output assembler commands to define the start of the
2638 constant pool for a function. FUNNAME is a string giving the name of the
2639 function. Should the return type of the function be required, it can be
2640 obtained via FUNDECL. SIZE is the size, in bytes, of the constant pool that
2641 will be written immediately after this call.
2643 If no constant-pool prefix is required, the usual case, this macro need not
2645 /* #define ASM_OUTPUT_POOL_PROLOGUE(FILE FUNNAME FUNDECL SIZE) */
2647 /* A C statement (with or without semicolon) to output a constant in the
2648 constant pool, if it needs special treatment. (This macro need not do
2649 anything for RTL expressions that can be output normally.)
2651 The argument FILE is the standard I/O stream to output the assembler code
2652 on. X is the RTL expression for the constant to output, and MODE is the
2653 machine mode (in case X is a `const_int'). ALIGN is the required alignment
2654 for the value X; you should output an assembler directive to force this much
2657 The argument LABELNO is a number to use in an internal label for the address
2658 of this pool entry. The definition of this macro is responsible for
2659 outputting the label definition at the proper place. Here is how to do
2662 ASM_OUTPUT_INTERNAL_LABEL (FILE, "LC", LABELNO);
2664 When you output a pool entry specially, you should end with a `goto' to the
2665 label JUMPTO. This will prevent the same pool entry from being output a
2666 second time in the usual manner.
2668 You need not define this macro if it would do nothing. */
2669 /* #define ASM_OUTPUT_SPECIAL_POOL_ENTRY(FILE, X, MODE, ALIGN, LABELNO, JUMPTO) */
2671 /* Define this macro as a C expression which is nonzero if the constant EXP, of
2672 type `tree', should be output after the code for a function. The compiler
2673 will normally output all constants before the function; you need not define
2674 this macro if this is OK. */
2675 /* #define CONSTANT_AFTER_FUNCTION_P(EXP) */
2677 /* A C statement to output assembler commands to at the end of the constant
2678 pool for a function. FUNNAME is a string giving the name of the function.
2679 Should the return type of the function be required, you can obtain it via
2680 FUNDECL. SIZE is the size, in bytes, of the constant pool that GNU CC wrote
2681 immediately before this call.
2683 If no constant-pool epilogue is required, the usual case, you need not
2684 define this macro. */
2685 /* #define ASM_OUTPUT_POOL_EPILOGUE (FILE FUNNAME FUNDECL SIZE) */
2687 /* Define this macro as a C expression which is nonzero if C is used as a
2688 logical line separator by the assembler.
2690 If you do not define this macro, the default is that only the character `;'
2691 is treated as a logical line separator. */
2692 #define IS_ASM_LOGICAL_LINE_SEPARATOR(C) ((C) == '|')
2695 /* Output of Uninitialized Variables. */
2697 /* A C statement (sans semicolon) to output to the stdio stream STREAM the
2698 assembler definition of a common-label named NAME whose size is SIZE bytes.
2699 The variable ROUNDED is the size rounded up to whatever alignment the caller
2702 Use the expression `assemble_name (STREAM, NAME)' to output the name itself;
2703 before and after that, output the additional assembler syntax for defining
2704 the name, and a newline.
2706 This macro controls how the assembler definitions of uninitialized global
2707 variables are output. */
2708 /* #define ASM_OUTPUT_COMMON(STREAM, NAME, SIZE, ROUNDED) */
2710 /* Like `ASM_OUTPUT_COMMON' except takes the required alignment as a separate,
2711 explicit argument. If you define this macro, it is used in place of
2712 `ASM_OUTPUT_COMMON', and gives you more flexibility in handling the required
2713 alignment of the variable. The alignment is specified as the number of
2716 Defined in svr4.h. */
2717 /* #define ASM_OUTPUT_ALIGNED_COMMON(STREAM, NAME, SIZE, ALIGNMENT) */
2719 /* Like ASM_OUTPUT_ALIGNED_COMMON except that it takes an additional argument -
2720 the DECL of the variable to be output, if there is one. This macro can be
2721 called with DECL == NULL_TREE. If you define this macro, it is used in
2722 place of both ASM_OUTPUT_COMMON and ASM_OUTPUT_ALIGNED_COMMON, and gives you
2723 more flexibility in handling the destination of the variable. */
2724 /* #define ASM_OUTPUT_DECL_COMMON (STREAM, DECL, NAME, SIZE, ALIGNMENT) */
2726 /* If defined, it is similar to `ASM_OUTPUT_COMMON', except that it is used
2727 when NAME is shared. If not defined, `ASM_OUTPUT_COMMON' will be used. */
2728 /* #define ASM_OUTPUT_SHARED_COMMON(STREAM, NAME, SIZE, ROUNDED) */
2730 /* A C statement (sans semicolon) to output to the stdio stream STREAM the
2731 assembler definition of uninitialized global DECL named NAME whose size is
2732 SIZE bytes. The variable ROUNDED is the size rounded up to whatever
2733 alignment the caller wants.
2735 Try to use function `asm_output_bss' defined in `varasm.c' when defining
2736 this macro. If unable, use the expression `assemble_name (STREAM, NAME)' to
2737 output the name itself; before and after that, output the additional
2738 assembler syntax for defining the name, and a newline.
2740 This macro controls how the assembler definitions of uninitialized global
2741 variables are output. This macro exists to properly support languages like
2742 `c++' which do not have `common' data. However, this macro currently is not
2743 defined for all targets. If this macro and `ASM_OUTPUT_ALIGNED_BSS' are not
2744 defined then `ASM_OUTPUT_COMMON' or `ASM_OUTPUT_ALIGNED_COMMON' or
2745 `ASM_OUTPUT_DECL_COMMON' is used. */
2746 /* #define ASM_OUTPUT_BSS(STREAM, DECL, NAME, SIZE, ROUNDED) */
2748 /* Like `ASM_OUTPUT_BSS' except takes the required alignment as a separate,
2749 explicit argument. If you define this macro, it is used in place of
2750 `ASM_OUTPUT_BSS', and gives you more flexibility in handling the required
2751 alignment of the variable. The alignment is specified as the number of
2754 Try to use function `asm_output_aligned_bss' defined in file `varasm.c' when
2755 defining this macro. */
2756 /* #define ASM_OUTPUT_ALIGNED_BSS(STREAM, DECL, NAME, SIZE, ALIGNMENT) */
2758 /* If defined, it is similar to `ASM_OUTPUT_BSS', except that it is used when
2759 NAME is shared. If not defined, `ASM_OUTPUT_BSS' will be used. */
2760 /* #define ASM_OUTPUT_SHARED_BSS(STREAM, DECL, NAME, SIZE, ROUNDED) */
2762 /* A C statement (sans semicolon) to output to the stdio stream STREAM the
2763 assembler definition of a local-common-label named NAME whose size is SIZE
2764 bytes. The variable ROUNDED is the size rounded up to whatever alignment
2767 Use the expression `assemble_name (STREAM, NAME)' to output the name itself;
2768 before and after that, output the additional assembler syntax for defining
2769 the name, and a newline.
2771 This macro controls how the assembler definitions of uninitialized static
2772 variables are output. */
2773 /* #define ASM_OUTPUT_LOCAL(STREAM, NAME, SIZE, ROUNDED) */
2775 /* Like `ASM_OUTPUT_LOCAL' except takes the required alignment as a separate,
2776 explicit argument. If you define this macro, it is used in place of
2777 `ASM_OUTPUT_LOCAL', and gives you more flexibility in handling the required
2778 alignment of the variable. The alignment is specified as the number of
2781 Defined in svr4.h. */
2782 /* #define ASM_OUTPUT_ALIGNED_LOCAL(STREAM, NAME, SIZE, ALIGNMENT) */
2784 /* Like `ASM_OUTPUT_ALIGNED_LOCAL' except that it takes an additional
2785 parameter - the DECL of variable to be output, if there is one.
2786 This macro can be called with DECL == NULL_TREE. If you define
2787 this macro, it is used in place of `ASM_OUTPUT_LOCAL' and
2788 `ASM_OUTPUT_ALIGNED_LOCAL', and gives you more flexibility in
2789 handling the destination of the variable. */
2790 /* #define ASM_OUTPUT_DECL_LOCAL(STREAM, DECL, NAME, SIZE, ALIGNMENT) */
2792 /* If defined, it is similar to `ASM_OUTPUT_LOCAL', except that it is used when
2793 NAME is shared. If not defined, `ASM_OUTPUT_LOCAL' will be used. */
2794 /* #define ASM_OUTPUT_SHARED_LOCAL (STREAM, NAME, SIZE, ROUNDED) */
2797 /* Output and Generation of Labels. */
2799 /* A C statement (sans semicolon) to output to the stdio stream STREAM the
2800 assembler definition of a label named NAME. Use the expression
2801 `assemble_name (STREAM, NAME)' to output the name itself; before and after
2802 that, output the additional assembler syntax for defining the name, and a
2804 #define ASM_OUTPUT_LABEL(STREAM, NAME) \
2806 assemble_name (STREAM, NAME); \
2807 fputs (":\n", STREAM); \
2810 /* A C statement to output to the stdio stream STREAM the assembler
2811 definition of a symbol named SYMBOL. */
2812 #define ASM_OUTPUT_SYMBOL_REF(STREAM, SYMBOL) \
2814 if (SYMBOL_REF_FLAG (SYMBOL)) \
2816 fputs ("@fptr(", STREAM); \
2817 assemble_name (STREAM, XSTR (SYMBOL, 0)); \
2818 fputc (')', STREAM); \
2821 assemble_name (STREAM, XSTR (SYMBOL, 0)); \
2824 /* A C statement to output to the stdio stream STREAM the assembler
2825 definition of a label, the textual form is in 'BUF'. Not used
2827 #define ASM_OUTPUT_LABEL_REF(STREAM, NAME) \
2829 fputs ("@fptr(", STREAM); \
2830 assemble_name (STREAM, NAME); \
2831 fputc (')', STREAM); \
2834 /* A C statement (sans semicolon) to output to the stdio stream STREAM any text
2835 necessary for declaring the name NAME of a function which is being defined.
2836 This macro is responsible for outputting the label definition (perhaps using
2837 `ASM_OUTPUT_LABEL'). The argument DECL is the `FUNCTION_DECL' tree node
2838 representing the function.
2840 If this macro is not defined, then the function name is defined in the usual
2841 manner as a label (by means of `ASM_OUTPUT_LABEL').
2843 Defined in svr4.h. */
2844 /* #define ASM_DECLARE_FUNCTION_NAME(STREAM, NAME, DECL) */
2846 /* A C statement (sans semicolon) to output to the stdio stream STREAM any text
2847 necessary for declaring the size of a function which is being defined. The
2848 argument NAME is the name of the function. The argument DECL is the
2849 `FUNCTION_DECL' tree node representing the function.
2851 If this macro is not defined, then the function size is not defined.
2853 Defined in svr4.h. */
2854 /* #define ASM_DECLARE_FUNCTION_SIZE(STREAM, NAME, DECL) */
2856 /* A C statement (sans semicolon) to output to the stdio stream STREAM any text
2857 necessary for declaring the name NAME of an initialized variable which is
2858 being defined. This macro must output the label definition (perhaps using
2859 `ASM_OUTPUT_LABEL'). The argument DECL is the `VAR_DECL' tree node
2860 representing the variable.
2862 If this macro is not defined, then the variable name is defined in the usual
2863 manner as a label (by means of `ASM_OUTPUT_LABEL').
2865 Defined in svr4.h. */
2866 /* #define ASM_DECLARE_OBJECT_NAME(STREAM, NAME, DECL) */
2868 /* A C statement (sans semicolon) to finish up declaring a variable name once
2869 the compiler has processed its initializer fully and thus has had a chance
2870 to determine the size of an array when controlled by an initializer. This
2871 is used on systems where it's necessary to declare something about the size
2874 If you don't define this macro, that is equivalent to defining it to do
2877 Defined in svr4.h. */
2878 /* #define ASM_FINISH_DECLARE_OBJECT(STREAM, DECL, TOPLEVEL, ATEND) */
2880 /* A C statement (sans semicolon) to output to the stdio stream STREAM some
2881 commands that will make the label NAME global; that is, available for
2882 reference from other files. Use the expression `assemble_name (STREAM,
2883 NAME)' to output the name itself; before and after that, output the
2884 additional assembler syntax for making that name global, and a newline. */
2885 #define ASM_GLOBALIZE_LABEL(STREAM,NAME) \
2887 fputs ("\t.globl ", STREAM); \
2888 assemble_name (STREAM, NAME); \
2889 fputs ("\n", STREAM); \
2892 /* A C statement (sans semicolon) to output to the stdio stream STREAM some
2893 commands that will make the label NAME weak; that is, available for
2894 reference from other files but only used if no other definition is
2895 available. Use the expression `assemble_name (STREAM, NAME)' to output the
2896 name itself; before and after that, output the additional assembler syntax
2897 for making that name weak, and a newline.
2899 If you don't define this macro, GNU CC will not support weak symbols and you
2900 should not define the `SUPPORTS_WEAK' macro.
2902 Defined in svr4.h. */
2903 /* #define ASM_WEAKEN_LABEL */
2905 /* A C expression which evaluates to true if the target supports weak symbols.
2907 If you don't define this macro, `defaults.h' provides a default definition.
2908 If `ASM_WEAKEN_LABEL' is defined, the default definition is `1'; otherwise,
2909 it is `0'. Define this macro if you want to control weak symbol support
2910 with a compiler flag such as `-melf'. */
2911 /* #define SUPPORTS_WEAK */
2913 /* A C statement (sans semicolon) to mark DECL to be emitted as a
2914 public symbol such that extra copies in multiple translation units
2915 will be discarded by the linker. Define this macro if your object
2916 file format provides support for this concept, such as the `COMDAT'
2917 section flags in the Microsoft Windows PE/COFF format, and this
2918 support requires changes to DECL, such as putting it in a separate
2921 Defined in svr4.h. */
2922 /* #define MAKE_DECL_ONE_ONLY */
2924 /* A C expression which evaluates to true if the target supports one-only
2927 If you don't define this macro, `varasm.c' provides a default definition.
2928 If `MAKE_DECL_ONE_ONLY' is defined, the default definition is `1';
2929 otherwise, it is `0'. Define this macro if you want to control one-only
2930 symbol support with a compiler flag, or if setting the `DECL_ONE_ONLY' flag
2931 is enough to mark a declaration to be emitted as one-only. */
2932 /* #define SUPPORTS_ONE_ONLY */
2934 /* A C statement (sans semicolon) to output to the stdio stream STREAM any text
2935 necessary for declaring the name of an external symbol named NAME which is
2936 referenced in this compilation but not defined. The value of DECL is the
2937 tree node for the declaration.
2939 This macro need not be defined if it does not need to output anything. The
2940 GNU assembler and most Unix assemblers don't require anything. */
2941 /* #define ASM_OUTPUT_EXTERNAL(STREAM, DECL, NAME) */
2943 /* A C statement (sans semicolon) to output on STREAM an assembler pseudo-op to
2944 declare a library function name external. The name of the library function
2945 is given by SYMREF, which has type `rtx' and is a `symbol_ref'.
2947 This macro need not be defined if it does not need to output anything. The
2948 GNU assembler and most Unix assemblers don't require anything.
2950 Defined in svr4.h. */
2951 /* #define ASM_OUTPUT_EXTERNAL_LIBCALL(STREAM, SYMREF) */
2953 /* A C statement (sans semicolon) to output to the stdio stream STREAM a
2954 reference in assembler syntax to a label named NAME. This should add `_' to
2955 the front of the name, if that is customary on your operating system, as it
2956 is in most Berkeley Unix systems. This macro is used in `assemble_name'. */
2957 /* #define ASM_OUTPUT_LABELREF(STREAM, NAME) */
2959 /* A C statement to output to the stdio stream STREAM a label whose name is
2960 made from the string PREFIX and the number NUM.
2962 It is absolutely essential that these labels be distinct from the labels
2963 used for user-level functions and variables. Otherwise, certain programs
2964 will have name conflicts with internal labels.
2966 It is desirable to exclude internal labels from the symbol table of the
2967 object file. Most assemblers have a naming convention for labels that
2968 should be excluded; on many systems, the letter `L' at the beginning of a
2969 label has this effect. You should find out what convention your system
2970 uses, and follow it.
2972 The usual definition of this macro is as follows:
2974 fprintf (STREAM, "L%s%d:\n", PREFIX, NUM)
2976 Defined in svr4.h. */
2977 /* #define ASM_OUTPUT_INTERNAL_LABEL(STREAM, PREFIX, NUM) */
2979 /* A C statement to store into the string STRING a label whose name is made
2980 from the string PREFIX and the number NUM.
2982 This string, when output subsequently by `assemble_name', should produce the
2983 output that `ASM_OUTPUT_INTERNAL_LABEL' would produce with the same PREFIX
2986 If the string begins with `*', then `assemble_name' will output the rest of
2987 the string unchanged. It is often convenient for
2988 `ASM_GENERATE_INTERNAL_LABEL' to use `*' in this way. If the string doesn't
2989 start with `*', then `ASM_OUTPUT_LABELREF' gets to output the string, and
2990 may change it. (Of course, `ASM_OUTPUT_LABELREF' is also part of your
2991 machine description, so you should know what it does on your machine.)
2993 Defined in svr4.h. */
2994 /* #define ASM_GENERATE_INTERNAL_LABEL(LABEL, PREFIX, NUM) */
2996 /* A C expression to assign to OUTVAR (which is a variable of type `char *') a
2997 newly allocated string made from the string NAME and the number NUMBER, with
2998 some suitable punctuation added. Use `alloca' to get space for the string.
3000 The string will be used as an argument to `ASM_OUTPUT_LABELREF' to produce
3001 an assembler label for an internal static variable whose name is NAME.
3002 Therefore, the string must be such as to result in valid assembler code.
3003 The argument NUMBER is different each time this macro is executed; it
3004 prevents conflicts between similarly-named internal static variables in
3007 Ideally this string should not be a valid C identifier, to prevent any
3008 conflict with the user's own symbols. Most assemblers allow periods or
3009 percent signs in assembler symbols; putting at least one of these between
3010 the name and the number will suffice. */
3011 #define ASM_FORMAT_PRIVATE_NAME(OUTVAR, NAME, NUMBER) \
3013 (OUTVAR) = (char *) alloca (strlen ((NAME)) + 12); \
3014 sprintf ((OUTVAR), "%s.%ld", (NAME), (long)(NUMBER)); \
3017 /* A C statement to output to the stdio stream STREAM assembler code which
3018 defines (equates) the symbol NAME to have the value VALUE.
3020 If SET_ASM_OP is defined, a default definition is provided which is correct
3023 Defined in svr4.h. */
3024 /* #define ASM_OUTPUT_DEF(STREAM, NAME, VALUE) */
3026 /* A C statement to output to the stdio stream STREAM assembler code which
3027 defines (equates) the weak symbol NAME to have the value VALUE.
3029 Define this macro if the target only supports weak aliases; define
3030 ASM_OUTPUT_DEF instead if possible. */
3031 /* #define ASM_OUTPUT_WEAK_ALIAS (STREAM, NAME, VALUE) */
3033 /* Define this macro to override the default assembler names used for Objective
3036 The default name is a unique method number followed by the name of the class
3037 (e.g. `_1_Foo'). For methods in categories, the name of the category is
3038 also included in the assembler name (e.g. `_1_Foo_Bar').
3040 These names are safe on most systems, but make debugging difficult since the
3041 method's selector is not present in the name. Therefore, particular systems
3042 define other ways of computing names.
3044 BUF is an expression of type `char *' which gives you a buffer in which to
3045 store the name; its length is as long as CLASS_NAME, CAT_NAME and SEL_NAME
3046 put together, plus 50 characters extra.
3048 The argument IS_INST specifies whether the method is an instance method or a
3049 class method; CLASS_NAME is the name of the class; CAT_NAME is the name of
3050 the category (or NULL if the method is not in a category); and SEL_NAME is
3051 the name of the selector.
3053 On systems where the assembler can handle quoted names, you can use this
3054 macro to provide more human-readable names. */
3055 /* #define OBJC_GEN_METHOD_LABEL(BUF, IS_INST, CLASS_NAME, CAT_NAME, SEL_NAME) */
3058 /* Macros Controlling Initialization Routines. */
3060 /* If defined, a C string constant for the assembler operation to identify the
3061 following data as initialization code. If not defined, GNU CC will assume
3062 such a section does not exist. When you are using special sections for
3063 initialization and termination functions, this macro also controls how
3064 `crtstuff.c' and `libgcc2.c' arrange to run the initialization functions.
3066 Defined in svr4.h. */
3067 /* #define INIT_SECTION_ASM_OP */
3069 /* If defined, `main' will not call `__main' as described above. This macro
3070 should be defined for systems that control the contents of the init section
3071 on a symbol-by-symbol basis, such as OSF/1, and should not be defined
3072 explicitly for systems that support `INIT_SECTION_ASM_OP'. */
3073 /* #define HAS_INIT_SECTION */
3075 /* If defined, a C string constant for a switch that tells the linker that the
3076 following symbol is an initialization routine. */
3077 /* #define LD_INIT_SWITCH */
3079 /* If defined, a C string constant for a switch that tells the linker that the
3080 following symbol is a finalization routine. */
3081 /* #define LD_FINI_SWITCH */
3083 /* If defined, `main' will call `__main' despite the presence of
3084 `INIT_SECTION_ASM_OP'. This macro should be defined for systems where the
3085 init section is not actually run automatically, but is still useful for
3086 collecting the lists of constructors and destructors. */
3087 /* #define INVOKE__main */
3089 /* Define this macro as a C statement to output on the stream STREAM the
3090 assembler code to arrange to call the function named NAME at initialization
3093 Assume that NAME is the name of a C function generated automatically by the
3094 compiler. This function takes no arguments. Use the function
3095 `assemble_name' to output the name NAME; this performs any system-specific
3096 syntactic transformations such as adding an underscore.
3098 If you don't define this macro, nothing special is output to arrange to call
3099 the function. This is correct when the function will be called in some
3100 other manner--for example, by means of the `collect2' program, which looks
3101 through the symbol table to find these functions by their names.
3103 Defined in svr4.h. */
3104 /* #define ASM_OUTPUT_CONSTRUCTOR(STREAM, NAME) */
3106 /* This is like `ASM_OUTPUT_CONSTRUCTOR' but used for termination functions
3107 rather than initialization functions.
3109 Defined in svr4.h. */
3110 /* #define ASM_OUTPUT_DESTRUCTOR(STREAM, NAME) */
3112 /* If your system uses `collect2' as the means of processing constructors, then
3113 that program normally uses `nm' to scan an object file for constructor
3114 functions to be called. On certain kinds of systems, you can define these
3115 macros to make `collect2' work faster (and, in some cases, make it work at
3118 /* Define this macro if the system uses COFF (Common Object File Format) object
3119 files, so that `collect2' can assume this format and scan object files
3120 directly for dynamic constructor/destructor functions. */
3121 /* #define OBJECT_FORMAT_COFF */
3123 /* Define this macro if the system uses ROSE format object files, so that
3124 `collect2' can assume this format and scan object files directly for dynamic
3125 constructor/destructor functions.
3127 These macros are effective only in a native compiler; `collect2' as
3128 part of a cross compiler always uses `nm' for the target machine. */
3129 /* #define OBJECT_FORMAT_ROSE */
3131 /* Define this macro if the system uses ELF format object files.
3133 Defined in svr4.h. */
3134 /* #define OBJECT_FORMAT_ELF */
3136 /* Define this macro as a C string constant containing the file name to use to
3137 execute `nm'. The default is to search the path normally for `nm'.
3139 If your system supports shared libraries and has a program to list the
3140 dynamic dependencies of a given library or executable, you can define these
3141 macros to enable support for running initialization and termination
3142 functions in shared libraries: */
3143 /* #define REAL_NM_FILE_NAME */
3145 /* Define this macro to a C string constant containing the name of the program
3146 which lists dynamic dependencies, like `"ldd"' under SunOS 4. */
3147 /* #define LDD_SUFFIX */
3149 /* Define this macro to be C code that extracts filenames from the output of
3150 the program denoted by `LDD_SUFFIX'. PTR is a variable of type `char *'
3151 that points to the beginning of a line of output from `LDD_SUFFIX'. If the
3152 line lists a dynamic dependency, the code must advance PTR to the beginning
3153 of the filename on that line. Otherwise, it must set PTR to `NULL'. */
3154 /* #define PARSE_LDD_OUTPUT (PTR) */
3157 /* Output of Assembler Instructions. */
3159 /* A C initializer containing the assembler's names for the machine registers,
3160 each one as a C string constant. This is what translates register numbers
3161 in the compiler into assembler language. */
3162 #define REGISTER_NAMES \
3163 { "r0", "r1", "r2", "r3", "r4", "r5", "r6", "r7", "r8", "r9", "r10", \
3164 "r11", "r12", "r13", "psw", "sp", "carry", "fp", "ap" }
3166 /* If defined, a C initializer for an array of structures containing a name and
3167 a register number. This macro defines additional names for hard registers,
3168 thus allowing the `asm' option in declarations to refer to registers using
3170 #define ADDITIONAL_REGISTER_NAMES \
3174 /* Define this macro if you are using an unusual assembler that requires
3175 different names for the machine instructions.
3177 The definition is a C statement or statements which output an assembler
3178 instruction opcode to the stdio stream STREAM. The macro-operand PTR is a
3179 variable of type `char *' which points to the opcode name in its "internal"
3180 form--the form that is written in the machine description. The definition
3181 should output the opcode name to STREAM, performing any translation you
3182 desire, and increment the variable PTR to point at the end of the opcode so
3183 that it will not be output twice.
3185 In fact, your macro definition may process less than the entire opcode name,
3186 or more than the opcode name; but if you want to process text that includes
3187 `%'-sequences to substitute operands, you must take care of the substitution
3188 yourself. Just be sure to increment PTR over whatever text should not be
3191 If you need to look at the operand values, they can be found as the elements
3192 of `recog_data.operand'.
3194 If the macro definition does nothing, the instruction is output in the usual
3196 /* #define ASM_OUTPUT_OPCODE(STREAM, PTR) */
3198 /* If defined, a C statement to be executed just prior to the output of
3199 assembler code for INSN, to modify the extracted operands so they will be
3202 Here the argument OPVEC is the vector containing the operands extracted from
3203 INSN, and NOPERANDS is the number of elements of the vector which contain
3204 meaningful data for this insn. The contents of this vector are what will be
3205 used to convert the insn template into assembler code, so you can change the
3206 assembler output by changing the contents of the vector.
3208 This macro is useful when various assembler syntaxes share a single file of
3209 instruction patterns; by defining this macro differently, you can cause a
3210 large class of instructions to be output differently (such as with
3211 rearranged operands). Naturally, variations in assembler syntax affecting
3212 individual insn patterns ought to be handled by writing conditional output
3213 routines in those patterns.
3215 If this macro is not defined, it is equivalent to a null statement. */
3216 /* #define FINAL_PRESCAN_INSN(INSN, OPVEC, NOPERANDS) */
3218 /* If defined, `FINAL_PRESCAN_INSN' will be called on each
3219 `CODE_LABEL'. In that case, OPVEC will be a null pointer and
3220 NOPERANDS will be zero. */
3221 /* #define FINAL_PRESCAN_LABEL */
3223 /* A C compound statement to output to stdio stream STREAM the assembler syntax
3224 for an instruction operand X. X is an RTL expression.
3226 CODE is a value that can be used to specify one of several ways of printing
3227 the operand. It is used when identical operands must be printed differently
3228 depending on the context. CODE comes from the `%' specification that was
3229 used to request printing of the operand. If the specification was just
3230 `%DIGIT' then CODE is 0; if the specification was `%LTR DIGIT' then CODE is
3231 the ASCII code for LTR.
3233 If X is a register, this macro should print the register's name. The names
3234 can be found in an array `reg_names' whose type is `char *[]'. `reg_names'
3235 is initialized from `REGISTER_NAMES'.
3237 When the machine description has a specification `%PUNCT' (a `%' followed by
3238 a punctuation character), this macro is called with a null pointer for X and
3239 the punctuation character for CODE. */
3240 #define PRINT_OPERAND(STREAM, X, CODE) xstormy16_print_operand (STREAM, X, CODE)
3242 /* A C expression which evaluates to true if CODE is a valid punctuation
3243 character for use in the `PRINT_OPERAND' macro. If
3244 `PRINT_OPERAND_PUNCT_VALID_P' is not defined, it means that no punctuation
3245 characters (except for the standard one, `%') are used in this way. */
3246 /* #define PRINT_OPERAND_PUNCT_VALID_P(CODE) */
3248 /* A C compound statement to output to stdio stream STREAM the assembler syntax
3249 for an instruction operand that is a memory reference whose address is X. X
3250 is an RTL expression. */
3251 #define PRINT_OPERAND_ADDRESS(STREAM, X) xstormy16_print_operand_address (STREAM, X)
3253 /* A C statement, to be executed after all slot-filler instructions have been
3254 output. If necessary, call `dbr_sequence_length' to determine the number of
3255 slots filled in a sequence (zero if not currently outputting a sequence), to
3256 decide how many no-ops to output, or whatever.
3258 Don't define this macro if it has nothing to do, but it is helpful in
3259 reading assembly output if the extent of the delay sequence is made explicit
3260 (e.g. with white space).
3262 Note that output routines for instructions with delay slots must be prepared
3263 to deal with not being output as part of a sequence (i.e. when the
3264 scheduling pass is not run, or when no slot fillers could be found.) The
3265 variable `final_sequence' is null when not processing a sequence, otherwise
3266 it contains the `sequence' rtx being output. */
3267 /* #define DBR_OUTPUT_SEQEND(FILE) */
3269 /* If defined, C string expressions to be used for the `%R', `%L', `%U', and
3270 `%I' options of `asm_fprintf' (see `final.c'). These are useful when a
3271 single `md' file must support multiple assembler formats. In that case, the
3272 various `tm.h' files can define these macros differently.
3274 USER_LABEL_PREFIX is defined in svr4.h. */
3275 #define REGISTER_PREFIX ""
3276 #define LOCAL_LABEL_PREFIX "."
3277 #define USER_LABEL_PREFIX ""
3278 #define IMMEDIATE_PREFIX "#"
3280 /* If your target supports multiple dialects of assembler language (such as
3281 different opcodes), define this macro as a C expression that gives the
3282 numeric index of the assembler language dialect to use, with zero as the
3285 If this macro is defined, you may use `{option0|option1|option2...}'
3286 constructs in the output templates of patterns or in the first argument of
3287 `asm_fprintf'. This construct outputs `option0', `option1' or `option2',
3288 etc., if the value of `ASSEMBLER_DIALECT' is zero, one or two, etc. Any
3289 special characters within these strings retain their usual meaning.
3291 If you do not define this macro, the characters `{', `|' and `}' do not have
3292 any special meaning when used in templates or operands to `asm_fprintf'.
3294 Define the macros `REGISTER_PREFIX', `LOCAL_LABEL_PREFIX',
3295 `USER_LABEL_PREFIX' and `IMMEDIATE_PREFIX' if you can express the variations
3296 in assemble language syntax with that mechanism. Define `ASSEMBLER_DIALECT'
3297 and use the `{option0|option1}' syntax if the syntax variant are larger and
3298 involve such things as different opcodes or operand order. */
3299 /* #define ASSEMBLER_DIALECT */
3301 /* A C expression to output to STREAM some assembler code which will push hard
3302 register number REGNO onto the stack. The code need not be optimal, since
3303 this macro is used only when profiling. */
3304 #define ASM_OUTPUT_REG_PUSH(STREAM, REGNO) \
3305 fprintf (STREAM, "\tpush %d\n", REGNO)
3307 /* A C expression to output to STREAM some assembler code which will pop hard
3308 register number REGNO off of the stack. The code need not be optimal, since
3309 this macro is used only when profiling. */
3310 #define ASM_OUTPUT_REG_POP(STREAM, REGNO) \
3311 fprintf (STREAM, "\tpop %d\n", REGNO)
3314 /* Output of dispatch tables. */
3316 /* This port does not use the ASM_OUTPUT_ADDR_VEC_ELT macro, because
3317 this could cause label alignment to appear between the 'br' and the table,
3318 which would be bad. Instead, it controls the output of the table
3320 #define ASM_OUTPUT_ADDR_VEC(LABEL, BODY) \
3321 xstormy16_output_addr_vec (file, LABEL, BODY)
3323 /* Alignment for ADDR_VECs is the same as for code. */
3324 #define ADDR_VEC_ALIGN(ADDR_VEC) 1
3327 /* Assembler Commands for Exception Regions. */
3329 /* An rtx used to mask the return address found via RETURN_ADDR_RTX, so that it
3330 does not contain any extraneous set bits in it. */
3331 /* #define MASK_RETURN_ADDR */
3333 /* Define this macro to 0 if your target supports DWARF 2 frame unwind
3334 information, but it does not yet work with exception handling. Otherwise,
3335 if your target supports this information (if it defines
3336 `INCOMING_RETURN_ADDR_RTX'), GCC will provide a default definition of 1.
3338 If this macro is defined to 1, the DWARF 2 unwinder will be the default
3339 exception handling mechanism; otherwise, setjmp/longjmp will be used by
3342 If this macro is defined to anything, the DWARF 2 unwinder will be used
3343 instead of inline unwinders and __unwind_function in the non-setjmp case. */
3344 #define DWARF2_UNWIND_INFO 0
3346 /* Don't use __builtin_setjmp for unwinding, since it's tricky to get
3347 at the high 16 bits of an address. */
3348 #define DONT_USE_BUILTIN_SETJMP
3349 #define JMP_BUF_SIZE 8
3351 /* Assembler Commands for Alignment. */
3353 /* The alignment (log base 2) to put in front of LABEL, which follows
3356 This macro need not be defined if you don't want any special alignment to be
3357 done at such a time. Most machine descriptions do not currently define the
3359 /* #define LABEL_ALIGN_AFTER_BARRIER(LABEL) */
3361 /* The desired alignment for the location counter at the beginning
3364 This macro need not be defined if you don't want any special alignment to be
3365 done at such a time. Most machine descriptions do not currently define the
3367 /* #define LOOP_ALIGN(LABEL) */
3369 /* A C statement to output to the stdio stream STREAM an assembler instruction
3370 to advance the location counter by NBYTES bytes. Those bytes should be zero
3371 when loaded. NBYTES will be a C expression of type `int'.
3373 Defined in elfos.h. */
3374 /* #define ASM_OUTPUT_SKIP(STREAM, NBYTES) */
3376 /* Define this macro if `ASM_OUTPUT_SKIP' should not be used in the text
3377 section because it fails put zeros in the bytes that are skipped. This is
3378 true on many Unix systems, where the pseudo-op to skip bytes produces no-op
3379 instructions rather than zeros when used in the text section. */
3380 /* #define ASM_NO_SKIP_IN_TEXT */
3382 /* A C statement to output to the stdio stream STREAM an assembler command to
3383 advance the location counter to a multiple of 2 to the POWER bytes. POWER
3384 will be a C expression of type `int'. */
3385 #define ASM_OUTPUT_ALIGN(STREAM, POWER) \
3386 fprintf ((STREAM), "\t.p2align %d\n", (POWER))
3389 /* Macros Affecting all Debug Formats. */
3391 /* A C expression that returns the integer offset value for an automatic
3392 variable having address X (an RTL expression). The default computation
3393 assumes that X is based on the frame-pointer and gives the offset from the
3394 frame-pointer. This is required for targets that produce debugging output
3395 for DBX or COFF-style debugging output for SDB and allow the frame-pointer
3396 to be eliminated when the `-g' options is used. */
3397 /* #define DEBUGGER_AUTO_OFFSET(X) */
3399 /* A C expression that returns the integer offset value for an argument having
3400 address X (an RTL expression). The nominal offset is OFFSET. */
3401 /* #define DEBUGGER_ARG_OFFSET(OFFSET, X) */
3403 /* A C expression that returns the type of debugging output GNU CC produces
3404 when the user specifies `-g' or `-ggdb'. Define this if you have arranged
3405 for GNU CC to support more than one format of debugging output. Currently,
3406 the allowable values are `DBX_DEBUG', `SDB_DEBUG', `DWARF_DEBUG',
3407 `DWARF2_DEBUG', and `XCOFF_DEBUG'.
3409 The value of this macro only affects the default debugging output; the user
3410 can always get a specific type of output by using `-gstabs', `-gcoff',
3411 `-gdwarf-1', `-gdwarf-2', or `-gxcoff'.
3413 Defined in svr4.h. */
3414 #undef PREFERRED_DEBUGGING_TYPE
3415 #define PREFERRED_DEBUGGING_TYPE DWARF2_DEBUG
3418 /* Specific Options for DBX Output. */
3420 /* Define this macro if GNU CC should produce debugging output for DBX in
3421 response to the `-g' option.
3423 Defined in svr4.h. */
3424 /* #define DBX_DEBUGGING_INFO */
3426 /* Define this macro if GNU CC should produce XCOFF format debugging output in
3427 response to the `-g' option. This is a variant of DBX format. */
3428 /* #define XCOFF_DEBUGGING_INFO */
3430 /* Define this macro to control whether GNU CC should by default generate GDB's
3431 extended version of DBX debugging information (assuming DBX-format debugging
3432 information is enabled at all). If you don't define the macro, the default
3433 is 1: always generate the extended information if there is any occasion to. */
3434 /* #define DEFAULT_GDB_EXTENSIONS */
3436 /* Define this macro if all `.stabs' commands should be output while in the
3438 /* #define DEBUG_SYMS_TEXT */
3440 /* A C string constant naming the assembler pseudo op to use instead of
3441 `.stabs' to define an ordinary debugging symbol. If you don't define this
3442 macro, `.stabs' is used. This macro applies only to DBX debugging
3443 information format. */
3444 /* #define ASM_STABS_OP */
3446 /* A C string constant naming the assembler pseudo op to use instead of
3447 `.stabd' to define a debugging symbol whose value is the current location.
3448 If you don't define this macro, `.stabd' is used. This macro applies only
3449 to DBX debugging information format. */
3450 /* #define ASM_STABD_OP */
3452 /* A C string constant naming the assembler pseudo op to use instead of
3453 `.stabn' to define a debugging symbol with no name. If you don't define
3454 this macro, `.stabn' is used. This macro applies only to DBX debugging
3455 information format. */
3456 /* #define ASM_STABN_OP */
3458 /* Define this macro if DBX on your system does not support the construct
3459 `xsTAGNAME'. On some systems, this construct is used to describe a forward
3460 reference to a structure named TAGNAME. On other systems, this construct is
3461 not supported at all. */
3462 /* #define DBX_NO_XREFS */
3464 /* A symbol name in DBX-format debugging information is normally continued
3465 (split into two separate `.stabs' directives) when it exceeds a certain
3466 length (by default, 80 characters). On some operating systems, DBX requires
3467 this splitting; on others, splitting must not be done. You can inhibit
3468 splitting by defining this macro with the value zero. You can override the
3469 default splitting-length by defining this macro as an expression for the
3470 length you desire. */
3471 /* #define DBX_CONTIN_LENGTH */
3473 /* Normally continuation is indicated by adding a `\' character to the end of a
3474 `.stabs' string when a continuation follows. To use a different character
3475 instead, define this macro as a character constant for the character you
3476 want to use. Do not define this macro if backslash is correct for your
3478 /* #define DBX_CONTIN_CHAR */
3480 /* Define this macro if it is necessary to go to the data section before
3481 outputting the `.stabs' pseudo-op for a non-global static variable. */
3482 /* #define DBX_STATIC_STAB_DATA_SECTION */
3484 /* The value to use in the "code" field of the `.stabs' directive for a
3485 typedef. The default is `N_LSYM'. */
3486 /* #define DBX_TYPE_DECL_STABS_CODE */
3488 /* The value to use in the "code" field of the `.stabs' directive for a static
3489 variable located in the text section. DBX format does not provide any
3490 "right" way to do this. The default is `N_FUN'. */
3491 /* #define DBX_STATIC_CONST_VAR_CODE */
3493 /* The value to use in the "code" field of the `.stabs' directive for a
3494 parameter passed in registers. DBX format does not provide any "right" way
3495 to do this. The default is `N_RSYM'. */
3496 /* #define DBX_REGPARM_STABS_CODE */
3498 /* The letter to use in DBX symbol data to identify a symbol as a parameter
3499 passed in registers. DBX format does not customarily provide any way to do
3500 this. The default is `'P''. */
3501 /* #define DBX_REGPARM_STABS_LETTER */
3503 /* The letter to use in DBX symbol data to identify a symbol as a stack
3504 parameter. The default is `'p''. */
3505 /* #define DBX_MEMPARM_STABS_LETTER */
3507 /* Define this macro if the DBX information for a function and its arguments
3508 should precede the assembler code for the function. Normally, in DBX
3509 format, the debugging information entirely follows the assembler code.
3511 Defined in svr4.h. */
3512 /* #define DBX_FUNCTION_FIRST */
3514 /* Define this macro if the `N_LBRAC' symbol for a block should precede the
3515 debugging information for variables and functions defined in that block.
3516 Normally, in DBX format, the `N_LBRAC' symbol comes first. */
3517 /* #define DBX_LBRAC_FIRST */
3519 /* Define this macro if the value of a symbol describing the scope of a block
3520 (`N_LBRAC' or `N_RBRAC') should be relative to the start of the enclosing
3521 function. Normally, GNU C uses an absolute address.
3523 Defined in svr4.h. */
3524 /* #define DBX_BLOCKS_FUNCTION_RELATIVE */
3526 /* Define this macro if GNU C should generate `N_BINCL' and `N_EINCL'
3527 stabs for included header files, as on Sun systems. This macro
3528 also directs GNU C to output a type number as a pair of a file
3529 number and a type number within the file. Normally, GNU C does not
3530 generate `N_BINCL' or `N_EINCL' stabs, and it outputs a single
3531 number for a type number. */
3532 /* #define DBX_USE_BINCL */
3535 /* Open ended Hooks for DBX Output. */
3537 /* Define this macro to say how to output to STREAM the debugging information
3538 for the start of a scope level for variable names. The argument NAME is the
3539 name of an assembler symbol (for use with `assemble_name') whose value is
3540 the address where the scope begins. */
3541 /* #define DBX_OUTPUT_LBRAC(STREAM, NAME) */
3543 /* Like `DBX_OUTPUT_LBRAC', but for the end of a scope level. */
3544 /* #define DBX_OUTPUT_RBRAC(STREAM, NAME) */
3546 /* Define this macro if the target machine requires special handling to output
3547 an enumeration type. The definition should be a C statement (sans
3548 semicolon) to output the appropriate information to STREAM for the type
3550 /* #define DBX_OUTPUT_ENUM(STREAM, TYPE) */
3552 /* Define this macro if the target machine requires special output at the end
3553 of the debugging information for a function. The definition should be a C
3554 statement (sans semicolon) to output the appropriate information to STREAM.
3555 FUNCTION is the `FUNCTION_DECL' node for the function. */
3556 /* #define DBX_OUTPUT_FUNCTION_END(STREAM, FUNCTION) */
3558 /* Define this macro if you need to control the order of output of the standard
3559 data types at the beginning of compilation. The argument SYMS is a `tree'
3560 which is a chain of all the predefined global symbols, including names of
3563 Normally, DBX output starts with definitions of the types for integers and
3564 characters, followed by all the other predefined types of the particular
3565 language in no particular order.
3567 On some machines, it is necessary to output different particular types
3568 first. To do this, define `DBX_OUTPUT_STANDARD_TYPES' to output those
3569 symbols in the necessary order. Any predefined types that you don't
3570 explicitly output will be output afterward in no particular order.
3572 Be careful not to define this macro so that it works only for C. There are
3573 no global variables to access most of the built-in types, because another
3574 language may have another set of types. The way to output a particular type
3575 is to look through SYMS to see if you can find it. Here is an example:
3579 for (decl = syms; decl; decl = TREE_CHAIN (decl))
3580 if (!strcmp (IDENTIFIER_POINTER (DECL_NAME (decl)),
3582 dbxout_symbol (decl);
3586 This does nothing if the expected type does not exist.
3588 See the function `init_decl_processing' in `c-decl.c' to find the names to
3589 use for all the built-in C types. */
3590 /* #define DBX_OUTPUT_STANDARD_TYPES(SYMS) */
3592 /* Some stabs encapsulation formats (in particular ECOFF), cannot
3593 handle the `.stabs "",N_FUN,,0,0,Lscope-function-1' gdb dbx
3594 extension construct. On those machines, define this macro to turn
3595 this feature off without disturbing the rest of the gdb extensions. */
3596 /* #define NO_DBX_FUNCTION_END */
3599 /* File names in DBX format. */
3601 /* Define this if DBX wants to have the current directory recorded in each
3604 Note that the working directory is always recorded if GDB extensions are
3606 /* #define DBX_WORKING_DIRECTORY */
3608 /* A C statement to output DBX debugging information to the stdio stream STREAM
3609 which indicates that file NAME is the main source file--the file specified
3610 as the input file for compilation. This macro is called only once, at the
3611 beginning of compilation.
3613 This macro need not be defined if the standard form of output for DBX
3614 debugging information is appropriate.
3616 Defined in svr4.h. */
3617 /* #define DBX_OUTPUT_MAIN_SOURCE_FILENAME(STREAM, NAME) */
3619 /* A C statement to output DBX debugging information to the stdio stream STREAM
3620 which indicates that the current directory during compilation is named NAME.
3622 This macro need not be defined if the standard form of output for DBX
3623 debugging information is appropriate. */
3624 /* #define DBX_OUTPUT_MAIN_SOURCE_DIRECTORY(STREAM, NAME) */
3626 /* A C statement to output DBX debugging information at the end of compilation
3627 of the main source file NAME.
3629 If you don't define this macro, nothing special is output at the end of
3630 compilation, which is correct for most machines. */
3631 /* #define DBX_OUTPUT_MAIN_SOURCE_FILE_END(STREAM, NAME) */
3633 /* A C statement to output DBX debugging information to the stdio stream STREAM
3634 which indicates that file NAME is the current source file. This output is
3635 generated each time input shifts to a different source file as a result of
3636 `#include', the end of an included file, or a `#line' command.
3638 This macro need not be defined if the standard form of output for DBX
3639 debugging information is appropriate. */
3640 /* #define DBX_OUTPUT_SOURCE_FILENAME(STREAM, NAME) */
3643 /* Macros for SDB and Dwarf Output. */
3645 /* Define this macro if GNU CC should produce COFF-style debugging output for
3646 SDB in response to the `-g' option. */
3647 /* #define SDB_DEBUGGING_INFO */
3649 /* Define this macro if GNU CC should produce dwarf format debugging output in
3650 response to the `-g' option.
3652 Defined in svr4.h. */
3653 /* #define DWARF_DEBUGGING_INFO */
3655 /* Define this macro if GNU CC should produce dwarf version 2 format debugging
3656 output in response to the `-g' option.
3658 To support optional call frame debugging information, you must also define
3659 `INCOMING_RETURN_ADDR_RTX' and either set `RTX_FRAME_RELATED_P' on the
3660 prologue insns if you use RTL for the prologue, or call `dwarf2out_def_cfa'
3661 and `dwarf2out_reg_save' as appropriate from `TARGET_ASM_FUNCTION_PROLOGUE'
3664 Defined in svr4.h. */
3665 /* #define DWARF2_DEBUGGING_INFO */
3667 /* Define this macro if GNU CC should produce dwarf version 2-style
3668 line numbers. This usually requires extending the assembler to
3669 support them, and #defining DWARF2_LINE_MIN_INSN_LENGTH in the
3670 assembler configuration header files. */
3671 /* #define DWARF2_ASM_LINE_DEBUG_INFO 1 */
3673 /* Define this macro if addresses in Dwarf 2 debugging info should not
3674 be the same size as pointers on the target architecture. The
3675 macro's value should be the size, in bytes, to use for addresses in
3678 Some architectures use word addresses to refer to code locations,
3679 but Dwarf 2 info always uses byte addresses. On such machines,
3680 Dwarf 2 addresses need to be larger than the architecture's
3682 #define DWARF2_ADDR_SIZE 4
3684 /* Define these macros to override the assembler syntax for the special SDB
3685 assembler directives. See `sdbout.c' for a list of these macros and their
3686 arguments. If the standard syntax is used, you need not define them
3688 /* #define PUT_SDB_... */
3690 /* Some assemblers do not support a semicolon as a delimiter, even between SDB
3691 assembler directives. In that case, define this macro to be the delimiter
3692 to use (usually `\n'). It is not necessary to define a new set of
3693 `PUT_SDB_OP' macros if this is the only change required. */
3694 /* #define SDB_DELIM */
3696 /* Define this macro to override the usual method of constructing a dummy name
3697 for anonymous structure and union types. See `sdbout.c' for more
3699 /* #define SDB_GENERATE_FAKE */
3701 /* Define this macro to allow references to unknown structure, union, or
3702 enumeration tags to be emitted. Standard COFF does not allow handling of
3703 unknown references, MIPS ECOFF has support for it. */
3704 /* #define SDB_ALLOW_UNKNOWN_REFERENCES */
3706 /* Define this macro to allow references to structure, union, or enumeration
3707 tags that have not yet been seen to be handled. Some assemblers choke if
3708 forward tags are used, while some require it. */
3709 /* #define SDB_ALLOW_FORWARD_REFERENCES */
3712 /* Miscellaneous Parameters. */
3714 /* Define this if you have defined special-purpose predicates in the file
3715 `MACHINE.c'. This macro is called within an initializer of an array of
3716 structures. The first field in the structure is the name of a predicate and
3717 the second field is an array of rtl codes. For each predicate, list all rtl
3718 codes that can be in expressions matched by the predicate. The list should
3719 have a trailing comma. Here is an example of two entries in the list for a
3720 typical RISC machine:
3722 #define PREDICATE_CODES \
3723 {"gen_reg_rtx_operand", {SUBREG, REG}}, \
3724 {"reg_or_short_cint_operand", {SUBREG, REG, CONST_INT}},
3726 Defining this macro does not affect the generated code (however, incorrect
3727 definitions that omit an rtl code that may be matched by the predicate can
3728 cause the compiler to malfunction). Instead, it allows the table built by
3729 `genrecog' to be more compact and efficient, thus speeding up the compiler.
3730 The most important predicates to include in the list specified by this macro
3731 are thoses used in the most insn patterns. */
3732 #define PREDICATE_CODES \
3733 {"shift_operator", {ASHIFT, ASHIFTRT, LSHIFTRT }}, \
3734 {"equality_operator", {EQ, NE }}, \
3735 {"inequality_operator", {GE, GT, LE, LT, GEU, GTU, LEU, LTU }}, \
3736 {"xstormy16_ineqsi_operator", {LT, GE, LTU, GEU }}, \
3737 {"nonimmediate_nonstack_operand", {REG, MEM}},
3738 /* An alias for a machine mode name. This is the machine mode that elements of
3739 a jump-table should have. */
3740 #define CASE_VECTOR_MODE SImode
3742 /* Define as C expression which evaluates to nonzero if the tablejump
3743 instruction expects the table to contain offsets from the address of the
3745 Do not define this if the table should contain absolute addresses. */
3746 /* #define CASE_VECTOR_PC_RELATIVE 1 */
3748 /* Define this if control falls through a `case' insn when the index value is
3749 out of range. This means the specified default-label is actually ignored by
3750 the `case' insn proper. */
3751 /* #define CASE_DROPS_THROUGH */
3753 /* Define this to be the smallest number of different values for which it is
3754 best to use a jump-table instead of a tree of conditional branches. The
3755 default is four for machines with a `casesi' instruction and five otherwise.
3756 This is best for most machines. */
3757 /* #define CASE_VALUES_THRESHOLD */
3759 /* Define this macro if operations between registers with integral mode smaller
3760 than a word are always performed on the entire register. Most RISC machines
3761 have this property and most CISC machines do not. */
3762 #define WORD_REGISTER_OPERATIONS
3764 /* Define this macro to be a C expression indicating when insns that read
3765 memory in MODE, an integral mode narrower than a word, set the bits outside
3766 of MODE to be either the sign-extension or the zero-extension of the data
3767 read. Return `SIGN_EXTEND' for values of MODE for which the insn
3768 sign-extends, `ZERO_EXTEND' for which it zero-extends, and `NIL' for other
3771 This macro is not called with MODE non-integral or with a width greater than
3772 or equal to `BITS_PER_WORD', so you may return any value in this case. Do
3773 not define this macro if it would always return `NIL'. On machines where
3774 this macro is defined, you will normally define it as the constant
3775 `SIGN_EXTEND' or `ZERO_EXTEND'. */
3776 #define LOAD_EXTEND_OP(MODE) ZERO_EXTEND
3778 /* Define if loading short immediate values into registers sign extends. */
3779 /* #define SHORT_IMMEDIATES_SIGN_EXTEND */
3781 /* Define this macro if the same instructions that convert a floating point
3782 number to a signed fixed point number also convert validly to an unsigned
3784 /* #define FIXUNS_TRUNC_LIKE_FIX_TRUNC */
3786 /* The maximum number of bytes that a single instruction can move quickly from
3787 memory to memory. */
3790 /* The maximum number of bytes that a single instruction can move quickly from
3791 memory to memory. If this is undefined, the default is `MOVE_MAX'.
3792 Otherwise, it is the constant value that is the largest value that
3793 `MOVE_MAX' can have at run-time. */
3794 /* #define MAX_MOVE_MAX */
3796 /* A C expression that is nonzero if on this machine the number of bits
3797 actually used for the count of a shift operation is equal to the number of
3798 bits needed to represent the size of the object being shifted. When this
3799 macro is non-zero, the compiler will assume that it is safe to omit a
3800 sign-extend, zero-extend, and certain bitwise `and' instructions that
3801 truncates the count of a shift operation. On machines that have
3802 instructions that act on bitfields at variable positions, which may include
3803 `bit test' instructions, a nonzero `SHIFT_COUNT_TRUNCATED' also enables
3804 deletion of truncations of the values that serve as arguments to bitfield
3807 If both types of instructions truncate the count (for shifts) and position
3808 (for bitfield operations), or if no variable-position bitfield instructions
3809 exist, you should define this macro.
3811 However, on some machines, such as the 80386 and the 680x0, truncation only
3812 applies to shift operations and not the (real or pretended) bitfield
3813 operations. Define `SHIFT_COUNT_TRUNCATED' to be zero on such machines.
3814 Instead, add patterns to the `md' file that include the implied truncation
3815 of the shift instructions.
3817 You need not define this macro if it would always have the value of zero. */
3818 #define SHIFT_COUNT_TRUNCATED 1
3820 /* A C expression which is nonzero if on this machine it is safe to "convert"
3821 an integer of INPREC bits to one of OUTPREC bits (where OUTPREC is smaller
3822 than INPREC) by merely operating on it as if it had only OUTPREC bits.
3824 On many machines, this expression can be 1.
3826 When `TRULY_NOOP_TRUNCATION' returns 1 for a pair of sizes for modes for
3827 which `MODES_TIEABLE_P' is 0, suboptimal code can result. If this is the
3828 case, making `TRULY_NOOP_TRUNCATION' return 0 in such cases may improve
3830 #define TRULY_NOOP_TRUNCATION(OUTPREC, INPREC) 1
3832 /* A C expression describing the value returned by a comparison operator with
3833 an integral mode and stored by a store-flag instruction (`sCOND') when the
3834 condition is true. This description must apply to *all* the `sCOND'
3835 patterns and all the comparison operators whose results have a `MODE_INT'
3838 A value of 1 or -1 means that the instruction implementing the comparison
3839 operator returns exactly 1 or -1 when the comparison is true and 0 when the
3840 comparison is false. Otherwise, the value indicates which bits of the
3841 result are guaranteed to be 1 when the comparison is true. This value is
3842 interpreted in the mode of the comparison operation, which is given by the
3843 mode of the first operand in the `sCOND' pattern. Either the low bit or the
3844 sign bit of `STORE_FLAG_VALUE' be on. Presently, only those bits are used
3847 If `STORE_FLAG_VALUE' is neither 1 or -1, the compiler will generate code
3848 that depends only on the specified bits. It can also replace comparison
3849 operators with equivalent operations if they cause the required bits to be
3850 set, even if the remaining bits are undefined. For example, on a machine
3851 whose comparison operators return an `SImode' value and where
3852 `STORE_FLAG_VALUE' is defined as `0x80000000', saying that just the sign bit
3853 is relevant, the expression
3855 (ne:SI (and:SI X (const_int POWER-OF-2)) (const_int 0))
3859 (ashift:SI X (const_int N))
3861 where N is the appropriate shift count to move the bit being tested into the
3864 There is no way to describe a machine that always sets the low-order bit for
3865 a true value, but does not guarantee the value of any other bits, but we do
3866 not know of any machine that has such an instruction. If you are trying to
3867 port GNU CC to such a machine, include an instruction to perform a
3868 logical-and of the result with 1 in the pattern for the comparison operators
3871 Often, a machine will have multiple instructions that obtain a value from a
3872 comparison (or the condition codes). Here are rules to guide the choice of
3873 value for `STORE_FLAG_VALUE', and hence the instructions to be used:
3875 * Use the shortest sequence that yields a valid definition for
3876 `STORE_FLAG_VALUE'. It is more efficient for the compiler to
3877 "normalize" the value (convert it to, e.g., 1 or 0) than for
3878 the comparison operators to do so because there may be
3879 opportunities to combine the normalization with other
3882 * For equal-length sequences, use a value of 1 or -1, with -1
3883 being slightly preferred on machines with expensive jumps and
3884 1 preferred on other machines.
3886 * As a second choice, choose a value of `0x80000001' if
3887 instructions exist that set both the sign and low-order bits
3888 but do not define the others.
3890 * Otherwise, use a value of `0x80000000'.
3892 Many machines can produce both the value chosen for `STORE_FLAG_VALUE' and
3893 its negation in the same number of instructions. On those machines, you
3894 should also define a pattern for those cases, e.g., one matching
3896 (set A (neg:M (ne:M B C)))
3898 Some machines can also perform `and' or `plus' operations on condition code
3899 values with less instructions than the corresponding `sCOND' insn followed
3900 by `and' or `plus'. On those machines, define the appropriate patterns.
3901 Use the names `incscc' and `decscc', respectively, for the the patterns
3902 which perform `plus' or `minus' operations on condition code values. See
3903 `rs6000.md' for some examples. The GNU Superoptizer can be used to find
3904 such instruction sequences on other machines.
3906 You need not define `STORE_FLAG_VALUE' if the machine has no store-flag
3908 /* #define STORE_FLAG_VALUE */
3910 /* A C expression that gives a non-zero floating point value that is returned
3911 when comparison operators with floating-point results are true. Define this
3912 macro on machine that have comparison operations that return floating-point
3913 values. If there are no such operations, do not define this macro. */
3914 /* #define FLOAT_STORE_FLAG_VALUE */
3916 /* An alias for the machine mode for pointers. On most machines, define this
3917 to be the integer mode corresponding to the width of a hardware pointer;
3918 `SImode' on 32-bit machine or `DImode' on 64-bit machines. On some machines
3919 you must define this to be one of the partial integer modes, such as
3922 The width of `Pmode' must be at least as large as the value of
3923 `POINTER_SIZE'. If it is not equal, you must define the macro
3924 `POINTERS_EXTEND_UNSIGNED' to specify how pointers are extended to `Pmode'. */
3925 #define Pmode HImode
3927 /* An alias for the machine mode used for memory references to functions being
3928 called, in `call' RTL expressions. On most machines this should be
3930 #define FUNCTION_MODE HImode
3932 /* A C expression for the maximum number of instructions above which the
3933 function DECL should not be inlined. DECL is a `FUNCTION_DECL' node.
3935 The default definition of this macro is 64 plus 8 times the number of
3936 arguments that the function accepts. Some people think a larger threshold
3937 should be used on RISC machines. */
3938 /* #define INTEGRATE_THRESHOLD(DECL) */
3940 /* Define this if the preprocessor should ignore `#sccs' directives and print
3943 Defined in svr4.h. */
3944 /* #define SCCS_DIRECTIVE */
3946 /* Define this macro if the system header files support C++ as well as C. This
3947 macro inhibits the usual method of using system header files in C++, which
3948 is to pretend that the file's contents are enclosed in `extern "C" {...}'. */
3949 #define NO_IMPLICIT_EXTERN_C
3951 /* Define this macro if you want to implement any pragmas. If defined, it
3952 should be a C expression to be executed when #pragma is seen. The
3953 argument GETC is a function which will return the next character in the
3954 input stream, or EOF if no characters are left. The argument UNGETC is
3955 a function which will push a character back into the input stream. The
3956 argument NAME is the word following #pragma in the input stream. The input
3957 stream pointer will be pointing just beyond the end of this word. The
3958 expression should return true if it handled the pragma, false otherwise.
3959 The input stream should be left undistrubed if false is returned, otherwise
3960 it should be pointing at the next character after the end of the pragma.
3961 Any characters left between the end of the pragma and the end of the line will
3964 It is generally a bad idea to implement new uses of `#pragma'. The only
3965 reason to define this macro is for compatibility with other compilers that
3966 do support `#pragma' for the sake of any user programs which already use it. */
3967 /* #define HANDLE_PRAGMA(GETC, UNGETC, NAME) handle_pragma (GETC, UNGETC, NAME) */
3969 /* Define this macro to handle System V style pragmas: #pragma pack and
3970 #pragma weak. Note, #pragma weak will only be supported if SUPPORT_WEAK is
3973 Defined in svr4.h. */
3974 #define HANDLE_SYSV_PRAGMA
3976 /* Define this macro if you want to support the Win32 style pragmas
3977 #pragma pack(push,<n>) and #pragma pack(pop). */
3978 /* HANDLE_PRAGMA_PACK_PUSH_POP 1 */
3980 /* Define this macro if the assembler does not accept the character `$' in
3981 label names. By default constructors and destructors in G++ have `$' in the
3982 identifiers. If this macro is defined, `.' is used instead.
3984 Defined in svr4.h. */
3985 /* #define NO_DOLLAR_IN_LABEL */
3987 /* Define this macro if the assembler does not accept the character `.' in
3988 label names. By default constructors and destructors in G++ have names that
3989 use `.'. If this macro is defined, these names are rewritten to avoid `.'. */
3990 /* #define NO_DOT_IN_LABEL */
3992 /* Define this macro if the target system expects every program's `main'
3993 function to return a standard "success" value by default (if no other value
3994 is explicitly returned).
3996 The definition should be a C statement (sans semicolon) to generate the
3997 appropriate rtl instructions. It is used only when compiling the end of
3999 /* #define DEFAULT_MAIN_RETURN */
4001 /* Define this if the target system supports the function `atexit' from the
4002 ANSI C standard. If this is not defined, and `INIT_SECTION_ASM_OP' is not
4003 defined, a default `exit' function will be provided to support C++.
4005 Defined by svr4.h */
4006 /* #define HAVE_ATEXIT */
4008 /* Define this if your `exit' function needs to do something besides calling an
4009 external function `_cleanup' before terminating with `_exit'. The
4010 `EXIT_BODY' macro is only needed if netiher `HAVE_ATEXIT' nor
4011 `INIT_SECTION_ASM_OP' are defined. */
4012 /* #define EXIT_BODY */
4014 /* Define this macro as a C expression that is nonzero if it is safe for the
4015 delay slot scheduler to place instructions in the delay slot of INSN, even
4016 if they appear to use a resource set or clobbered in INSN. INSN is always a
4017 `jump_insn' or an `insn'; GNU CC knows that every `call_insn' has this
4018 behavior. On machines where some `insn' or `jump_insn' is really a function
4019 call and hence has this behavior, you should define this macro.
4021 You need not define this macro if it would always return zero. */
4022 /* #define INSN_SETS_ARE_DELAYED(INSN) */
4024 /* Define this macro as a C expression that is nonzero if it is safe for the
4025 delay slot scheduler to place instructions in the delay slot of INSN, even
4026 if they appear to set or clobber a resource referenced in INSN. INSN is
4027 always a `jump_insn' or an `insn'. On machines where some `insn' or
4028 `jump_insn' is really a function call and its operands are registers whose
4029 use is actually in the subroutine it calls, you should define this macro.
4030 Doing so allows the delay slot scheduler to move instructions which copy
4031 arguments into the argument registers into the delay slot of INSN.
4033 You need not define this macro if it would always return zero. */
4034 /* #define INSN_REFERENCES_ARE_DELAYED(INSN) */
4036 /* In rare cases, correct code generation requires extra machine dependent
4037 processing between the second jump optimization pass and delayed branch
4038 scheduling. On those machines, define this macro as a C statement to act on
4039 the code starting at INSN. */
4040 /* #define MACHINE_DEPENDENT_REORG(INSN) */
4042 /* Define this macro if in some cases global symbols from one translation unit
4043 may not be bound to undefined symbols in another translation unit without
4044 user intervention. For instance, under Microsoft Windows symbols must be
4045 explicitly imported from shared libraries (DLLs). */
4046 /* #define MULTIPLE_SYMBOL_SPACES */
4048 /* A C expression for the maximum number of instructions to execute via
4049 conditional execution instructions instead of a branch. A value of
4050 BRANCH_COST+1 is the default if the machine does not use
4051 cc0, and 1 if it does use cc0. */
4052 /* #define MAX_CONDITIONAL_EXECUTE */
4054 /* A C statement that adds to tree CLOBBERS a set of STRING_CST trees for any
4055 hard regs the port wishes to automatically clobber for all asms. */
4056 /* #define MD_ASM_CLOBBERS(CLOBBERS) */
4058 /* Indicate how many instructions can be issued at the same time. */
4059 /* #define ISSUE_RATE */
4061 /* A C statement which is executed by the Haifa scheduler at the beginning of
4062 each block of instructions that are to be scheduled. FILE is either a null
4063 pointer, or a stdio stream to write any debug output to. VERBOSE is the
4064 verbose level provided by -fsched-verbose-<n>. */
4065 /* #define MD_SCHED_INIT (FILE, VERBOSE) */
4067 /* A C statement which is executed by the Haifa scheduler after it has scheduled
4068 the ready list to allow the machine description to reorder it (for example to
4069 combine two small instructions together on VLIW machines). FILE is either a
4070 null pointer, or a stdio stream to write any debug output to. VERBOSE is the
4071 verbose level provided by -fsched-verbose-=<n>. READY is a pointer to the
4072 ready list of instructions that are ready to be scheduled. N_READY is the
4073 number of elements in the ready list. The scheduler reads the ready list in
4074 reverse order, starting with READY[N_READY-1] and going to READY[0]. CLOCK
4075 is the timer tick of the scheduler. CAN_ISSUE_MORE is an output parameter that
4076 is set to the number of insns that can issue this clock; normally this is just
4078 /* #define MD_SCHED_REORDER (FILE, VERBOSE, READY, N_READY, CLOCK, CAN_ISSUE_MORE) */
4080 /* A C statement which is executed by the Haifa scheduler after it has scheduled
4081 an insn from the ready list. FILE is either a null pointer, or a stdio stream
4082 to write any debug output to. VERBOSE is the verbose level provided by
4083 -fsched-verbose-<n>. INSN is the instruction that was scheduled. MORE is the
4084 number of instructions that can be issued in the current cycle. This macro
4085 is responsible for updating the value of MORE (typically by (MORE)--). */
4086 /* #define MD_SCHED_VARIABLE_ISSUE (FILE, VERBOSE, INSN, MORE) */
4088 /* Define this to the largest integer machine mode which can be used for
4089 operations other than load, store and copy operations. You need only define
4090 this macro if the target holds values larger than word_mode in general purpose
4091 registers. Most targets should not define this macro. */
4092 /* #define MAX_INTEGER_COMPUTATION_MODE */
4094 /* Define this macro as a C string constant for the linker argument to link in the
4095 system math library, or "" if the target does not have a separate math library.
4096 You need only define this macro if the default of "-lm" is wrong. */
4097 /* #define MATH_LIBRARY */
4099 /* Define the information needed to generate branch and scc insns. This is
4100 stored from the compare operation. Note that we can't use "rtx" here
4101 since it hasn't been defined! */
4103 extern struct rtx_def *xstormy16_compare_op0, *xstormy16_compare_op1;
4105 /* End of xstormy16.h */
OSDN Git Service
