1/* $NetBSD: bcopy_page.S,v 1.5 2002/08/17 16:36:33 thorpej Exp $ */ 2 3/* 4 * Copyright (c) 1995 Scott Stevens 5 * All rights reserved. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 10 * 1. Redistributions of source code must retain the above copyright 11 * notice, this list of conditions and the following disclaimer. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice, this list of conditions and the following disclaimer in the 14 * documentation and/or other materials provided with the distribution. 15 * 3. All advertising materials mentioning features or use of this software 16 * must display the following acknowledgement: 17 * This product includes software developed by Scott Stevens. 18 * 4. The name of the author may not be used to endorse or promote products 19 * derived from this software without specific prior written permission. 20 * 21 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR 22 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES 23 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. 24 * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, 25 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT 26 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 27 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 28 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 29 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF 30 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 31 * 32 * RiscBSD kernel project 33 * 34 * bcopy_page.S 35 * 36 * page optimised bcopy and bzero routines 37 * 38 * Created : 08/04/95 39 */ 40 41#include <machine/param.h> 42#include <machine/asm.h> 43 44/* #define BIG_LOOPS */ 45 46/* 47 * bcopy_page(src, dest) 48 * 49 * Optimised copy page routine. 50 * 51 * On entry: 52 * r0 - src address 53 * r1 - dest address 54 * 55 * Requires: 56 * number of bytes per page (NBPG) is a multiple of 512 (BIG_LOOPS), 128 57 * otherwise. 58 */ 59 60#define CHUNK_SIZE 32 61 62#ifdef __XSCALE__ 63 /* Conveniently, the chunk size is the XScale cache line size. */ 64#define PREFETCH_FIRST_CHUNK pld [r0] 65#define PREFETCH_NEXT_CHUNK pld [r0, #(CHUNK_SIZE)] 66#else 67#define PREFETCH_FIRST_CHUNK /* nothing */ 68#define PREFETCH_NEXT_CHUNK /* nothing */ 69#endif 70 71#ifndef COPY_CHUNK 72#define COPY_CHUNK \ 73 PREFETCH_NEXT_CHUNK ; \ 74 ldmia r0!, {r3-r8,ip,lr} ; \ 75 stmia r1!, {r3-r8,ip,lr} 76#endif /* ! COPY_CHUNK */ 77 78#ifndef SAVE_REGS 79#define SAVE_REGS stmfd sp!, {r4-r8, lr} 80#define RESTORE_REGS ldmfd sp!, {r4-r8, pc} 81#endif 82 83ENTRY(bcopy_page) 84 PREFETCH_FIRST_CHUNK 85 SAVE_REGS 86#ifdef BIG_LOOPS 87 mov r2, #(NBPG >> 9) 88#else 89 mov r2, #(NBPG >> 7) 90#endif 91 921: 93 COPY_CHUNK 94 COPY_CHUNK 95 COPY_CHUNK 96 COPY_CHUNK 97 98#ifdef BIG_LOOPS 99 /* There is little point making the loop any larger; unless we are 100 running with the cache off, the load/store overheads will 101 completely dominate this loop. */ 102 COPY_CHUNK 103 COPY_CHUNK 104 COPY_CHUNK 105 COPY_CHUNK 106 107 COPY_CHUNK 108 COPY_CHUNK 109 COPY_CHUNK 110 COPY_CHUNK 111 112 COPY_CHUNK 113 COPY_CHUNK 114 COPY_CHUNK 115 COPY_CHUNK 116#endif 117 subs r2, r2, #1 118 bne 1b 119 120 RESTORE_REGS /* ...and return. */ 121 122/* 123 * bzero_page(dest) 124 * 125 * Optimised zero page routine. 126 * 127 * On entry: 128 * r0 - dest address 129 * 130 * Requires: 131 * number of bytes per page (NBPG) is a multiple of 512 (BIG_LOOPS), 128 132 * otherwise 133 */ 134 135ENTRY(bzero_page) 136 stmfd sp!, {r4-r8, lr} 137#ifdef BIG_LOOPS 138 mov r2, #(NBPG >> 9) 139#else 140 mov r2, #(NBPG >> 7) 141#endif 142 mov r3, #0 143 mov r4, #0 144 mov r5, #0 145 mov r6, #0 146 mov r7, #0 147 mov r8, #0 148 mov ip, #0 149 mov lr, #0 150 1511: 152 stmia r0!, {r3-r8,ip,lr} 153 stmia r0!, {r3-r8,ip,lr} 154 stmia r0!, {r3-r8,ip,lr} 155 stmia r0!, {r3-r8,ip,lr} 156 157#ifdef BIG_LOOPS 158 /* There is little point making the loop any larger; unless we are 159 running with the cache off, the load/store overheads will 160 completely dominate this loop. */ 161 stmia r0!, {r3-r8,ip,lr} 162 stmia r0!, {r3-r8,ip,lr} 163 stmia r0!, {r3-r8,ip,lr} 164 stmia r0!, {r3-r8,ip,lr} 165 166 stmia r0!, {r3-r8,ip,lr} 167 stmia r0!, {r3-r8,ip,lr} 168 stmia r0!, {r3-r8,ip,lr} 169 stmia r0!, {r3-r8,ip,lr} 170 171 stmia r0!, {r3-r8,ip,lr} 172 stmia r0!, {r3-r8,ip,lr} 173 stmia r0!, {r3-r8,ip,lr} 174 stmia r0!, {r3-r8,ip,lr} 175 176#endif 177 178 subs r2, r2, #1 179 bne 1b 180 181 ldmfd sp!, {r4-r8, pc} 182