diff options
Diffstat (limited to 'lib/libc/mips/string/bzero.S')
| -rw-r--r-- | lib/libc/mips/string/bzero.S | 69 |
1 files changed, 44 insertions, 25 deletions
diff --git a/lib/libc/mips/string/bzero.S b/lib/libc/mips/string/bzero.S index 9b341302f975..66f29ddeeae8 100644 --- a/lib/libc/mips/string/bzero.S +++ b/lib/libc/mips/string/bzero.S @@ -1,4 +1,4 @@ -/* $NetBSD: bzero.S,v 1.8 2003/08/07 16:42:16 agc Exp $ */ +/* $NetBSD: bzero.S,v 1.10 2009/12/14 02:53:52 matt Exp $ */ /*- * Copyright (c) 1991, 1993 @@ -36,10 +36,15 @@ __FBSDID("$FreeBSD$"); #if defined(LIBC_SCCS) && !defined(lint) +#if 0 ASMSTR("from: @(#)bzero.s 8.1 (Berkeley) 6/4/93") - ASMSTR("$NetBSD: bzero.S,v 1.8 2003/08/07 16:42:16 agc Exp $") +#else + ASMSTR("$NetBSD: bzero.S,v 1.10 2009/12/14 02:53:52 matt Exp $") +#endif #endif /* LIBC_SCCS and not lint */ +#define _LOCORE /* XXX not really, just assembly-code source */ +#include <machine/endian.h> #ifdef __ABICALLS__ .abicalls @@ -49,34 +54,48 @@ __FBSDID("$FreeBSD$"); LEAF(bzero) .set noreorder - blt a1, 12, smallclr # small amount to clear? - subu a3, zero, a0 # compute # bytes to word align address - and a3, a3, 3 - beq a3, zero, 1f # skip if word aligned - subu a1, a1, a3 # subtract from remaining count -#ifdef __MIPSEB__ - swl zero, 0(a0) # clear 1, 2, or 3 bytes to align -#else - swr zero, 0(a0) # clear 1, 2, or 3 bytes to align + blt a1, 3*SZREG, smallclr # small amount to clear? + PTR_SUBU a3, zero, a0 # compute # bytes to word align address + and a3, a3, SZREG-1 + beq a3, zero, 1f # skip if word aligned +#if SZREG == 4 + PTR_SUBU a1, a1, a3 # subtract from remaining count + SWHI zero, 0(a0) # clear 1, 2, or 3 bytes to align + PTR_ADDU a0, a0, a3 +#endif +#if SZREG == 8 + PTR_SUBU a1, a1, a3 # subtract from remaining count + PTR_ADDU a0, a0, a3 # align dst to next word + sll a3, a3, 3 # bits to bytes + li a2, -1 # make a mask +#if _BYTE_ORDER == _BIG_ENDIAN + REG_SRLV a2, a2, a3 # we want to keep the MSB bytes +#endif +#if _BYTE_ORDER == _LITTLE_ENDIAN + REG_SLLV a2, a2, a3 # we want to keep the LSB bytes +#endif + nor a2, zero, a2 # complement the mask + REG_L v0, -SZREG(a0) # load the word to partially clear + and v0, v0, a2 # clear the bytes + REG_S v0, -SZREG(a0) # store it back #endif - addu a0, a0, a3 1: - and v0, a1, 3 # compute number of words left - subu a3, a1, v0 - move a1, v0 - addu a3, a3, a0 # compute ending address + and v0, a1, SZREG-1 # compute number of words left + PTR_SUBU a3, a1, v0 + move a1, v0 + PTR_ADDU a3, a3, a0 # compute ending address 2: - addu a0, a0, 4 # clear words - bne a0, a3, 2b # unrolling loop doesnt help - sw zero, -4(a0) # since we are limited by memory speed + PTR_ADDU a0, a0, SZREG # clear words + bne a0, a3, 2b # unrolling loop doesnt help + REG_S zero, -SZREG(a0) # since we are limited by memory speed smallclr: - ble a1, zero, 2f - addu a3, a1, a0 # compute ending address + ble a1, zero, 2f + PTR_ADDU a3, a1, a0 # compute ending address 1: - addu a0, a0, 1 # clear bytes - bne a0, a3, 1b - sb zero, -1(a0) + PTR_ADDU a0, a0, 1 # clear bytes + bne a0, a3, 1b + sb zero, -1(a0) 2: - j ra + j ra nop END(bzero) |