lib1funcs.asm

linux下的gcc编译器

/* libgcc routines for NEC V850.
   Copyright (C) 1996, 1997, 2002 Free Software Foundation, Inc.

This file is part of GNU CC.

GNU CC is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by the
Free Software Foundation; either version 2, or (at your option) any
later version.

In addition to the permissions in the GNU General Public License, the
Free Software Foundation gives you unlimited permission to link the
compiled version of this file into combinations with other programs,
and to distribute those combinations without any restriction coming
from the use of this file.  (The General Public License restrictions
do apply in other respects; for example, they cover modification of
the file, and distribution when not linked into a combine
executable.)

This file is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
General Public License for more details.

You should have received a copy of the GNU General Public License
along with this program; see the file COPYING.  If not, write to
the Free Software Foundation, 59 Temple Place - Suite 330,
Boston, MA 02111-1307, USA.  */

#ifdef L_mulsi3
	.text
	.globl ___mulsi3
	.type  ___mulsi3,@function
___mulsi3:
#ifdef __v850__	
/*
   #define SHIFT 12
   #define MASK ((1 << SHIFT) - 1)
    
   #define STEP(i, j)                               \
   ({                                               \
       short a_part = (a >> (i)) & MASK;            \
       short b_part = (b >> (j)) & MASK;            \
       int res = (((int) a_part) * ((int) b_part)); \
       res;                                         \
   })
  
   int
   __mulsi3 (unsigned a, unsigned b)
   {
      return STEP (0, 0) +
          ((STEP (SHIFT, 0) + STEP (0, SHIFT)) << SHIFT) +
          ((STEP (0, 2 * SHIFT) + STEP (SHIFT, SHIFT) + STEP (2 * SHIFT, 0))
           << (2 * SHIFT));
   }
*/
        mov   r6, r14
        movea lo(32767), r0, r10
        and   r10, r14
        mov   r7,  r15
        and   r10, r15
        shr   15,  r6
        mov   r6,  r13
        and   r10, r13
        shr   15,  r7
        mov   r7,  r12
        and   r10, r12
        shr   15,  r6
        shr   15,  r7
        mov   r14, r10
        mulh  r15, r10
        mov   r14, r11
        mulh  r12, r11
        mov   r13, r16
        mulh  r15, r16
        mulh  r14, r7
        mulh  r15, r6
        add   r16, r11
        mulh  r13, r12
        shl   15,  r11
        add   r11, r10
        add   r12, r7
        add   r6,  r7
        shl   30,  r7
        add   r7,  r10
        jmp   [r31]
#endif /* __v850__ */
#if defined(__v850e__) || defined(__v850ea__)
        /* This routine is almost unneccesarry because gcc
           generates the MUL instruction for the RTX mulsi3.
           But if someone wants to link his application with
           previsously compiled v850 objects then they will 
	   need this function.  */
 
        /* It isn't good to put the inst sequence as below;
              mul r7, r6,
              mov r6, r10, r0
           In this case, there is a RAW hazard between them.
           MUL inst takes 2 cycle in EX stage, then MOV inst
           must wait 1cycle.  */
        mov   r7, r10
        mul   r6, r10, r0
        jmp   [r31]
#endif /* __v850e__ */
	.size ___mulsi3,.-___mulsi3
#endif /* L_mulsi3 */


#ifdef L_udivsi3
	.text
	.global ___udivsi3
	.type	___udivsi3,@function
___udivsi3:
#ifdef __v850__
	mov 1,r12
	mov 0,r10
	cmp r6,r7
	bnl .L12
	movhi hi(-2147483648),r0,r13
	cmp r0,r7
	blt .L12
.L4:
	shl 1,r7
	shl 1,r12
	cmp r6,r7
	bnl .L12
	cmp r0,r12
	be .L8
	mov r7,r19
	and r13,r19
	be .L4
	br .L12
.L9:
	cmp r7,r6
	bl .L10
	sub r7,r6
	or r12,r10
.L10:
	shr 1,r12
	shr 1,r7
.L12:
	cmp r0,r12
	bne .L9
.L8:
	jmp [r31]

#else /* defined(__v850e__) */

	/* See comments at end of __mulsi3.  */
	mov   r6, r10	
	divu  r7, r10, r0
	jmp   [r31]		

#endif /* __v850e__ */

	.size ___udivsi3,.-___udivsi3
#endif

#ifdef L_divsi3
	.text
	.globl ___divsi3
	.type  ___divsi3,@function
___divsi3:
#ifdef __v850__
	add -8,sp
	st.w r31,4[sp]
	st.w r22,0[sp]
	mov 1,r22
	tst r7,r7
	bp .L3
	subr r0,r7
	subr r0,r22
.L3:
	tst r6,r6
	bp .L4
	subr r0,r6
	subr r0,r22
.L4:
	jarl ___udivsi3,r31
	cmp r0,r22
	bp .L7
	subr r0,r10
.L7:
	ld.w 0[sp],r22
	ld.w 4[sp],r31
	add 8,sp
	jmp [r31]

#else /* defined(__v850e__) */

	/* See comments at end of __mulsi3.  */
	mov   r6, r10
	div   r7, r10, r0
	jmp   [r31]

#endif /* __v850e__ */

	.size ___divsi3,.-___divsi3
#endif

#ifdef  L_umodsi3
	.text
	.globl ___umodsi3
	.type  ___umodsi3,@function
___umodsi3:
#ifdef __v850__
	add -12,sp
	st.w r31,8[sp]
	st.w r7,4[sp]
	st.w r6,0[sp]
	jarl ___udivsi3,r31
	ld.w 4[sp],r7
	mov r10,r6
	jarl ___mulsi3,r31
	ld.w 0[sp],r6
	subr r6,r10
	ld.w 8[sp],r31
	add 12,sp
	jmp [r31]

#else /* defined(__v850e__) */

	/* See comments at end of __mulsi3.  */
	divu  r7, r6, r10
	jmp   [r31]

#endif /* __v850e__ */

	.size ___umodsi3,.-___umodsi3
#endif /* L_umodsi3 */

#ifdef  L_modsi3
	.text
	.globl ___modsi3
	.type  ___modsi3,@function
___modsi3:
#ifdef __v850__	
	add -12,sp
	st.w r31,8[sp]
	st.w r7,4[sp]
	st.w r6,0[sp]
	jarl ___divsi3,r31
	ld.w 4[sp],r7
	mov r10,r6
	jarl ___mulsi3,r31
	ld.w 0[sp],r6
	subr r6,r10
	ld.w 8[sp],r31
	add 12,sp
	jmp [r31]

#else /* defined(__v850e__) */

	/* See comments at end of __mulsi3.  */
	div  r7, r6, r10
	jmp [r31]

#endif /* __v850e__ */

	.size ___modsi3,.-___modsi3
#endif /* L_modsi3 */

#ifdef	L_save_2
	.text
	.align	2
	.globl	__save_r2_r29
	.type	__save_r2_r29,@function
	/* Allocate space and save registers 2, 20 .. 29 on the stack */
	/* Called via:	jalr __save_r2_r29,r10 */
__save_r2_r29:
	mov	ep,r1
	addi	-44,sp,sp
	mov	sp,ep
	sst.w	r29,0[ep]
	sst.w	r28,4[ep]
	sst.w	r27,8[ep]
	sst.w	r26,12[ep]
	sst.w	r25,16[ep]
	sst.w	r24,20[ep]
	sst.w	r23,24[ep]
	sst.w	r22,28[ep]
	sst.w	r21,32[ep]
	sst.w	r20,36[ep]
	sst.w	r2,40[ep]
	mov	r1,ep
	jmp	[r10]
	.size	__save_r2_r29,.-__save_r2_r29

	/* Restore saved registers, deallocate stack and return to the user */
	/* Called via:	jr __return_r2_r29 */
	.align	2
	.globl	__return_r2_r29
	.type	__return_r2_r29,@function
__return_r2_r29:
	mov	ep,r1
	mov	sp,ep
	sld.w	0[ep],r29
	sld.w	4[ep],r28
	sld.w	8[ep],r27
	sld.w	12[ep],r26
	sld.w	16[ep],r25
	sld.w	20[ep],r24
	sld.w	24[ep],r23
	sld.w	28[ep],r22
	sld.w	32[ep],r21
	sld.w	36[ep],r20
	sld.w	40[ep],r2
	addi	44,sp,sp
	mov	r1,ep
	jmp	[r31]
	.size	__return_r2_r29,.-__return_r2_r29
#endif /* L_save_2 */

#ifdef	L_save_20
	.text
	.align	2
	.globl	__save_r20_r29
	.type	__save_r20_r29,@function
	/* Allocate space and save registers 20 .. 29 on the stack */
	/* Called via:	jalr __save_r20_r29,r10 */
__save_r20_r29:
	mov	ep,r1
	addi	-40,sp,sp
	mov	sp,ep
	sst.w	r29,0[ep]
	sst.w	r28,4[ep]
	sst.w	r27,8[ep]
	sst.w	r26,12[ep]
	sst.w	r25,16[ep]
	sst.w	r24,20[ep]
	sst.w	r23,24[ep]
	sst.w	r22,28[ep]
	sst.w	r21,32[ep]
	sst.w	r20,36[ep]
	mov	r1,ep
	jmp	[r10]
	.size	__save_r20_r29,.-__save_r20_r29

	/* Restore saved registers, deallocate stack and return to the user */
	/* Called via:	jr __return_r20_r29 */
	.align	2
	.globl	__return_r20_r29
	.type	__return_r20_r29,@function
__return_r20_r29:
	mov	ep,r1
	mov	sp,ep
	sld.w	0[ep],r29
	sld.w	4[ep],r28
	sld.w	8[ep],r27
	sld.w	12[ep],r26
	sld.w	16[ep],r25
	sld.w	20[ep],r24
	sld.w	24[ep],r23
	sld.w	28[ep],r22
	sld.w	32[ep],r21
	sld.w	36[ep],r20
	addi	40,sp,sp
	mov	r1,ep
	jmp	[r31]
	.size	__return_r20_r29,.-__return_r20_r29
#endif /* L_save_20 */

#ifdef	L_save_21
	.text
	.align	2
	.globl	__save_r21_r29
	.type	__save_r21_r29,@function
	/* Allocate space and save registers 21 .. 29 on the stack */
	/* Called via:	jalr __save_r21_r29,r10 */
__save_r21_r29:
	mov	ep,r1
	addi	-36,sp,sp
	mov	sp,ep
	sst.w	r29,0[ep]
	sst.w	r28,4[ep]
	sst.w	r27,8[ep]
	sst.w	r26,12[ep]
	sst.w	r25,16[ep]
	sst.w	r24,20[ep]
	sst.w	r23,24[ep]
	sst.w	r22,28[ep]
	sst.w	r21,32[ep]
	mov	r1,ep
	jmp	[r10]
	.size	__save_r21_r29,.-__save_r21_r29

	/* Restore saved registers, deallocate stack and return to the user */
	/* Called via:	jr __return_r21_r29 */
	.align	2
	.globl	__return_r21_r29
	.type	__return_r21_r29,@function
__return_r21_r29:
	mov	ep,r1
	mov	sp,ep
	sld.w	0[ep],r29
	sld.w	4[ep],r28
	sld.w	8[ep],r27
	sld.w	12[ep],r26
	sld.w	16[ep],r25
	sld.w	20[ep],r24
	sld.w	24[ep],r23
	sld.w	28[ep],r22
	sld.w	32[ep],r21
	addi	36,sp,sp
	mov	r1,ep
	jmp	[r31]
	.size	__return_r21_r29,.-__return_r21_r29
#endif /* L_save_21 */

#ifdef	L_save_22
	.text
	.align	2
	.globl	__save_r22_r29
	.type	__save_r22_r29,@function
	/* Allocate space and save registers 22 .. 29 on the stack */
	/* Called via:	jalr __save_r22_r29,r10 */
__save_r22_r29:
	mov	ep,r1
	addi	-32,sp,sp
	mov	sp,ep
	sst.w	r29,0[ep]
	sst.w	r28,4[ep]
	sst.w	r27,8[ep]
	sst.w	r26,12[ep]
	sst.w	r25,16[ep]
	sst.w	r24,20[ep]
	sst.w	r23,24[ep]
	sst.w	r22,28[ep]
	mov	r1,ep
	jmp	[r10]
	.size	__save_r22_r29,.-__save_r22_r29

	/* Restore saved registers, deallocate stack and return to the user */
	/* Called via:	jr __return_r22_r29 */
	.align	2
	.globl	__return_r22_r29
	.type	__return_r22_r29,@function
__return_r22_r29:
	mov	ep,r1
	mov	sp,ep
	sld.w	0[ep],r29
	sld.w	4[ep],r28
	sld.w	8[ep],r27
	sld.w	12[ep],r26
	sld.w	16[ep],r25
	sld.w	20[ep],r24
	sld.w	24[ep],r23
	sld.w	28[ep],r22
	addi	32,sp,sp
	mov	r1,ep
	jmp	[r31]
	.size	__return_r22_r29,.-__return_r22_r29
#endif /* L_save_22 */

#ifdef	L_save_23
	.text
	.align	2
	.globl	__save_r23_r29
	.type	__save_r23_r29,@function
	/* Allocate space and save registers 23 .. 29 on the stack */
	/* Called via:	jalr __save_r23_r29,r10 */
__save_r23_r29:
	mov	ep,r1
	addi	-28,sp,sp
	mov	sp,ep
	sst.w	r29,0[ep]
	sst.w	r28,4[ep]
	sst.w	r27,8[ep]
	sst.w	r26,12[ep]
	sst.w	r25,16[ep]
	sst.w	r24,20[ep]
	sst.w	r23,24[ep]
	mov	r1,ep
	jmp	[r10]
	.size	__save_r23_r29,.-__save_r23_r29

	/* Restore saved registers, deallocate stack and return to the user */
	/* Called via:	jr __return_r23_r29 */
	.align	2
	.globl	__return_r23_r29
	.type	__return_r23_r29,@function
__return_r23_r29:
	mov	ep,r1
	mov	sp,ep
	sld.w	0[ep],r29
	sld.w	4[ep],r28
	sld.w	8[ep],r27
	sld.w	12[ep],r26
	sld.w	16[ep],r25
	sld.w	20[ep],r24
	sld.w	24[ep],r23
	addi	28,sp,sp
	mov	r1,ep
	jmp	[r31]
	.size	__return_r23_r29,.-__return_r23_r29
#endif /* L_save_23 */

#ifdef	L_save_24
	.text
	.align	2
	.globl	__save_r24_r29
	.type	__save_r24_r29,@function
	/* Allocate space and save registers 24 .. 29 on the stack */
	/* Called via:	jalr __save_r24_r29,r10 */
__save_r24_r29:
	mov	ep,r1
	addi	-24,sp,sp
	mov	sp,ep
	sst.w	r29,0[ep]
	sst.w	r28,4[ep]
	sst.w	r27,8[ep]
	sst.w	r26,12[ep]
	sst.w	r25,16[ep]
	sst.w	r24,20[ep]
	mov	r1,ep
	jmp	[r10]
	.size	__save_r24_r29,.-__save_r24_r29

	/* Restore saved registers, deallocate stack and return to the user */
	/* Called via:	jr __return_r24_r29 */
	.align	2
	.globl	__return_r24_r29
	.type	__return_r24_r29,@function
__return_r24_r29:
	mov	ep,r1
	mov	sp,ep
	sld.w	0[ep],r29
	sld.w	4[ep],r28
	sld.w	8[ep],r27
	sld.w	12[ep],r26
	sld.w	16[ep],r25
	sld.w	20[ep],r24
	addi	24,sp,sp
	mov	r1,ep
	jmp	[r31]
	.size	__return_r24_r29,.-__return_r24_r29
#endif /* L_save_24 */

#ifdef	L_save_25
	.text
	.align	2
	.globl	__save_r25_r29
	.type	__save_r25_r29,@function
	/* Allocate space and save registers 25 .. 29 on the stack */
	/* Called via:	jalr __save_r25_r29,r10 */
__save_r25_r29:
	mov	ep,r1
	addi	-20,sp,sp
	mov	sp,ep
	sst.w	r29,0[ep]
	sst.w	r28,4[ep]
	sst.w	r27,8[ep]
	sst.w	r26,12[ep]
	sst.w	r25,16[ep]
	mov	r1,ep
	jmp	[r10]
	.size	__save_r25_r29,.-__save_r25_r29

	/* Restore saved registers, deallocate stack and return to the user */
	/* Called via:	jr __return_r25_r29 */
	.align	2
	.globl	__return_r25_r29
	.type	__return_r25_r29,@function
__return_r25_r29:
	mov	ep,r1
	mov	sp,ep
	sld.w	0[ep],r29
	sld.w	4[ep],r28
	sld.w	8[ep],r27
	sld.w	12[ep],r26
	sld.w	16[ep],r25
	addi	20,sp,sp
	mov	r1,ep
	jmp	[r31]
	.size	__return_r25_r29,.-__return_r25_r29
#endif /* L_save_25 */

#ifdef	L_save_26
	.text
	.align	2
	.globl	__save_r26_r29
	.type	__save_r26_r29,@function
	/* Allocate space and save registers 26 .. 29 on the stack */
	/* Called via:	jalr __save_r26_r29,r10 */
__save_r26_r29:
	mov	ep,r1
	add	-16,sp
	mov	sp,ep
	sst.w	r29,0[ep]
	sst.w	r28,4[ep]
	sst.w	r27,8[ep]
	sst.w	r26,12[ep]
	mov	r1,ep
	jmp	[r10]
	.size	__save_r26_r29,.-__save_r26_r29

	/* Restore saved registers, deallocate stack and return to the user */
	/* Called via:	jr __return_r26_r29 */
	.align	2
	.globl	__return_r26_r29
	.type	__return_r26_r29,@function
__return_r26_r29:
	mov	ep,r1
	mov	sp,ep
	sld.w	0[ep],r29
	sld.w	4[ep],r28
	sld.w	8[ep],r27
	sld.w	12[ep],r26
	addi	16,sp,sp
	mov	r1,ep
	jmp	[r31]
	.size	__return_r26_r29,.-__return_r26_r29
#endif /* L_save_26 */

#ifdef	L_save_27
	.text
	.align	2
	.globl	__save_r27_r29
	.type	__save_r27_r29,@function
	/* Allocate space and save registers 27 .. 29 on the stack */
	/* Called via:	jalr __save_r27_r29,r10 */
__save_r27_r29:
	add	-12,sp
	st.w	r29,0[sp]
	st.w	r28,4[sp]
	st.w	r27,8[sp]
	jmp	[r10]
	.size	__save_r27_r29,.-__save_r27_r29