a14ee64402
operation (by subtracting the absolute result from 0), don't test for overflow. This avoids an arithmetic exception when dividing LONG_MIN by 1: This is the only case that causes overflow, and the resulting value is correct under 2's compliment arithmetic. PR: 72024 Approved by: dwmalone@ Obtained from: NetBSD MFC after: 4 days
199 lines
4.7 KiB
Plaintext
199 lines
4.7 KiB
Plaintext
/* $NetBSD: divrem.m4,v 1.7 1996/10/17 03:08:04 cgd Exp $ */
|
|
/* $FreeBSD$ */
|
|
|
|
/*
|
|
* Copyright (c) 1994, 1995 Carnegie-Mellon University.
|
|
* All rights reserved.
|
|
*
|
|
* Author: Chris G. Demetriou
|
|
*
|
|
* Permission to use, copy, modify and distribute this software and
|
|
* its documentation is hereby granted, provided that both the copyright
|
|
* notice and this permission notice appear in all copies of the
|
|
* software, derivative works or modified versions, and any portions
|
|
* thereof, and that both notices appear in supporting documentation.
|
|
*
|
|
* CARNEGIE MELLON ALLOWS FREE USE OF THIS SOFTWARE IN ITS "AS IS"
|
|
* CONDITION. CARNEGIE MELLON DISCLAIMS ANY LIABILITY OF ANY KIND
|
|
* FOR ANY DAMAGES WHATSOEVER RESULTING FROM THE USE OF THIS SOFTWARE.
|
|
*
|
|
* Carnegie Mellon requests users of this software to return to
|
|
*
|
|
* Software Distribution Coordinator or Software.Distribution@CS.CMU.EDU
|
|
* School of Computer Science
|
|
* Carnegie Mellon University
|
|
* Pittsburgh PA 15213-3890
|
|
*
|
|
* any improvements or extensions that they make and grant Carnegie the
|
|
* rights to redistribute these changes.
|
|
*/
|
|
|
|
/*
|
|
* Division and remainder.
|
|
*
|
|
* The use of m4 is modeled after the sparc code, but the algorithm is
|
|
* simple binary long division.
|
|
*
|
|
* Note that the loops could probably benefit from unrolling.
|
|
*/
|
|
|
|
/*
|
|
* M4 Parameters
|
|
* NAME name of function to generate
|
|
* OP OP=div: t10 / t11 -> t12; OP=rem: t10 % t11 -> t12
|
|
* S S=true: signed; S=false: unsigned
|
|
* WORDSIZE total number of bits
|
|
*/
|
|
|
|
define(A, `t10')
|
|
define(B, `t11')
|
|
define(RESULT, `t12')
|
|
|
|
define(BIT, `t0')
|
|
define(I, `t1')
|
|
define(CC, `t2')
|
|
define(T_0, `t3')
|
|
ifelse(S, `true', `define(NEG, `t4')')
|
|
|
|
#include <machine/asm.h>
|
|
|
|
NESTED(NAME, 0, 0, t9, 0, 0) /* Get the right ra */
|
|
lda sp, -64(sp)
|
|
stq BIT, 0(sp)
|
|
stq I, 8(sp)
|
|
stq CC, 16(sp)
|
|
stq T_0, 24(sp)
|
|
ifelse(S, `true',
|
|
` stq NEG, 32(sp)')
|
|
stq A, 40(sp)
|
|
stq B, 48(sp)
|
|
mov zero, RESULT /* Initialize result to zero */
|
|
|
|
ifelse(S, `true',
|
|
`
|
|
/* Compute sign of result. If either is negative, this is easy. */
|
|
or A, B, NEG /* not the sign, but... */
|
|
srl NEG, WORDSIZE - 1, NEG /* rather, or of high bits */
|
|
blbc NEG, Ldoit /* neither negative? do it! */
|
|
|
|
ifelse(OP, `div',
|
|
` xor A, B, NEG /* THIS is the sign! */
|
|
', ` mov A, NEG /* sign follows A. */
|
|
')
|
|
srl NEG, WORDSIZE - 1, NEG /* make negation the low bit. */
|
|
|
|
srl A, WORDSIZE - 1, I /* is A negative? */
|
|
blbc I, LnegB /* no. */
|
|
/* A is negative; flip it. */
|
|
ifelse(WORDSIZE, `32', `
|
|
/* top 32 bits may be random junk */
|
|
zap A, 0xf0, A
|
|
')
|
|
subq zero, A, A
|
|
srl B, WORDSIZE - 1, I /* is B negative? */
|
|
blbc I, Ldoit /* no. */
|
|
LnegB:
|
|
/* B is definitely negative, no matter how we got here. */
|
|
ifelse(WORDSIZE, `32', `
|
|
/* top 32 bits may be random junk */
|
|
zap B, 0xf0, B
|
|
')
|
|
subq zero, B, B
|
|
Ldoit:
|
|
')
|
|
ifelse(WORDSIZE, `32', `
|
|
/*
|
|
* Clear the top 32 bits of each operand, as they may
|
|
* sign extension (if negated above), or random junk.
|
|
*/
|
|
zap A, 0xf0, A
|
|
zap B, 0xf0, B
|
|
')
|
|
|
|
/* kill the special cases. */
|
|
beq B, Ldotrap /* division by zero! */
|
|
|
|
cmpult A, B, CC /* A < B? */
|
|
/* RESULT is already zero, from above. A is untouched. */
|
|
bne CC, Lret_result
|
|
|
|
cmpeq A, B, CC /* A == B? */
|
|
cmovne CC, 1, RESULT
|
|
cmovne CC, zero, A
|
|
bne CC, Lret_result
|
|
|
|
/*
|
|
* Find out how many bits of zeros are at the beginning of the divisor.
|
|
*/
|
|
LBbits:
|
|
ldiq T_0, 1 /* I = 0; BIT = 1<<WORDSIZE-1 */
|
|
mov zero, I
|
|
sll T_0, WORDSIZE-1, BIT
|
|
LBloop:
|
|
and B, BIT, CC /* if bit in B is set, done. */
|
|
bne CC, LAbits
|
|
addq I, 1, I /* increment I, shift bit */
|
|
srl BIT, 1, BIT
|
|
cmplt I, WORDSIZE-1, CC /* if I leaves one bit, done. */
|
|
bne CC, LBloop
|
|
|
|
LAbits:
|
|
beq I, Ldodiv /* If I = 0, divide now. */
|
|
ldiq T_0, 1 /* BIT = 1<<WORDSIZE-1 */
|
|
sll T_0, WORDSIZE-1, BIT
|
|
|
|
LAloop:
|
|
and A, BIT, CC /* if bit in A is set, done. */
|
|
bne CC, Ldodiv
|
|
subq I, 1, I /* decrement I, shift bit */
|
|
srl BIT, 1, BIT
|
|
bne I, LAloop /* If I != 0, loop again */
|
|
|
|
Ldodiv:
|
|
sll B, I, B /* B <<= i */
|
|
ldiq T_0, 1
|
|
sll T_0, I, BIT
|
|
|
|
Ldivloop:
|
|
cmpult A, B, CC
|
|
or RESULT, BIT, T_0
|
|
cmoveq CC, T_0, RESULT
|
|
subq A, B, T_0
|
|
cmoveq CC, T_0, A
|
|
srl BIT, 1, BIT
|
|
srl B, 1, B
|
|
beq A, Lret_result
|
|
bne BIT, Ldivloop
|
|
|
|
Lret_result:
|
|
ifelse(OP, `div',
|
|
`', ` mov A, RESULT
|
|
')
|
|
ifelse(S, `true',
|
|
`
|
|
/* Check to see if we should negate it. */
|
|
subq zero, RESULT, T_0
|
|
cmovlbs NEG, T_0, RESULT
|
|
')
|
|
|
|
ldq BIT, 0(sp)
|
|
ldq I, 8(sp)
|
|
ldq CC, 16(sp)
|
|
ldq T_0, 24(sp)
|
|
ifelse(S, `true',
|
|
` ldq NEG, 32(sp)')
|
|
ldq A, 40(sp)
|
|
ldq B, 48(sp)
|
|
lda sp, 64(sp)
|
|
ret zero, (t9), 1
|
|
|
|
Ldotrap:
|
|
ldiq a0, -2 /* This is the signal to SIGFPE! */
|
|
call_pal PAL_gentrap
|
|
ifelse(OP, `div',
|
|
`', ` mov zero, A /* so that zero will be returned */
|
|
')
|
|
br zero, Lret_result
|
|
|
|
END(NAME)
|