bcopy_page.S revision 1.6.2.3
1 /* $NetBSD: bcopy_page.S,v 1.6.2.3 2004/09/21 13:13:08 skrll 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/asm.h> 42 43 #include "assym.h" 44 45 #ifndef __XSCALE__ 46 47 /* #define BIG_LOOPS */ 48 49 /* 50 * bcopy_page(src, dest) 51 * 52 * Optimised copy page routine. 53 * 54 * On entry: 55 * r0 - src address 56 * r1 - dest address 57 * 58 * Requires: 59 * number of bytes per page (PAGE_SIZE) is a multiple of 512 (BIG_LOOPS), 128 60 * otherwise. 61 */ 62 63 #define CHUNK_SIZE 32 64 65 #define PREFETCH_FIRST_CHUNK /* nothing */ 66 #define PREFETCH_NEXT_CHUNK /* nothing */ 67 68 #ifndef COPY_CHUNK 69 #define COPY_CHUNK \ 70 PREFETCH_NEXT_CHUNK ; \ 71 ldmia r0!, {r3-r8,ip,lr} ; \ 72 stmia r1!, {r3-r8,ip,lr} 73 #endif /* ! COPY_CHUNK */ 74 75 #ifndef SAVE_REGS 76 #define SAVE_REGS stmfd sp!, {r4-r8, lr} 77 #define RESTORE_REGS ldmfd sp!, {r4-r8, pc} 78 #endif 79 80 ENTRY(bcopy_page) 81 PREFETCH_FIRST_CHUNK 82 SAVE_REGS 83 #ifdef BIG_LOOPS 84 mov r2, #(PAGE_SIZE >> 9) 85 #else 86 mov r2, #(PAGE_SIZE >> 7) 87 #endif 88 89 1: 90 COPY_CHUNK 91 COPY_CHUNK 92 COPY_CHUNK 93 COPY_CHUNK 94 95 #ifdef BIG_LOOPS 96 /* There is little point making the loop any larger; unless we are 97 running with the cache off, the load/store overheads will 98 completely dominate this loop. */ 99 COPY_CHUNK 100 COPY_CHUNK 101 COPY_CHUNK 102 COPY_CHUNK 103 104 COPY_CHUNK 105 COPY_CHUNK 106 COPY_CHUNK 107 COPY_CHUNK 108 109 COPY_CHUNK 110 COPY_CHUNK 111 COPY_CHUNK 112 COPY_CHUNK 113 #endif 114 subs r2, r2, #1 115 bne 1b 116 117 RESTORE_REGS /* ...and return. */ 118 119 /* 120 * bzero_page(dest) 121 * 122 * Optimised zero page routine. 123 * 124 * On entry: 125 * r0 - dest address 126 * 127 * Requires: 128 * number of bytes per page (PAGE_SIZE) is a multiple of 512 (BIG_LOOPS), 128 129 * otherwise 130 */ 131 132 ENTRY(bzero_page) 133 stmfd sp!, {r4-r8, lr} 134 #ifdef BIG_LOOPS 135 mov r2, #(PAGE_SIZE >> 9) 136 #else 137 mov r2, #(PAGE_SIZE >> 7) 138 #endif 139 mov r3, #0 140 mov r4, #0 141 mov r5, #0 142 mov r6, #0 143 mov r7, #0 144 mov r8, #0 145 mov ip, #0 146 mov lr, #0 147 148 1: 149 stmia r0!, {r3-r8,ip,lr} 150 stmia r0!, {r3-r8,ip,lr} 151 stmia r0!, {r3-r8,ip,lr} 152 stmia r0!, {r3-r8,ip,lr} 153 154 #ifdef BIG_LOOPS 155 /* There is little point making the loop any larger; unless we are 156 running with the cache off, the load/store overheads will 157 completely dominate this loop. */ 158 stmia r0!, {r3-r8,ip,lr} 159 stmia r0!, {r3-r8,ip,lr} 160 stmia r0!, {r3-r8,ip,lr} 161 stmia r0!, {r3-r8,ip,lr} 162 163 stmia r0!, {r3-r8,ip,lr} 164 stmia r0!, {r3-r8,ip,lr} 165 stmia r0!, {r3-r8,ip,lr} 166 stmia r0!, {r3-r8,ip,lr} 167 168 stmia r0!, {r3-r8,ip,lr} 169 stmia r0!, {r3-r8,ip,lr} 170 stmia r0!, {r3-r8,ip,lr} 171 stmia r0!, {r3-r8,ip,lr} 172 173 #endif 174 175 subs r2, r2, #1 176 bne 1b 177 178 ldmfd sp!, {r4-r8, pc} 179 180 #else /* __XSCALE__ */ 181 182 /* 183 * XSCALE version of bcopy_page 184 */ 185 ENTRY(bcopy_page) 186 pld [r0] 187 stmfd sp!, {r4, r5} 188 mov ip, #32 189 ldr r2, [r0], #0x04 /* 0x00 */ 190 ldr r3, [r0], #0x04 /* 0x04 */ 191 1: pld [r0, #0x18] /* Prefetch 0x20 */ 192 ldr r4, [r0], #0x04 /* 0x08 */ 193 ldr r5, [r0], #0x04 /* 0x0c */ 194 strd r2, [r1], #0x08 195 ldr r2, [r0], #0x04 /* 0x10 */ 196 ldr r3, [r0], #0x04 /* 0x14 */ 197 strd r4, [r1], #0x08 198 ldr r4, [r0], #0x04 /* 0x18 */ 199 ldr r5, [r0], #0x04 /* 0x1c */ 200 strd r2, [r1], #0x08 201 ldr r2, [r0], #0x04 /* 0x20 */ 202 ldr r3, [r0], #0x04 /* 0x24 */ 203 pld [r0, #0x18] /* Prefetch 0x40 */ 204 strd r4, [r1], #0x08 205 ldr r4, [r0], #0x04 /* 0x28 */ 206 ldr r5, [r0], #0x04 /* 0x2c */ 207 strd r2, [r1], #0x08 208 ldr r2, [r0], #0x04 /* 0x30 */ 209 ldr r3, [r0], #0x04 /* 0x34 */ 210 strd r4, [r1], #0x08 211 ldr r4, [r0], #0x04 /* 0x38 */ 212 ldr r5, [r0], #0x04 /* 0x3c */ 213 strd r2, [r1], #0x08 214 ldr r2, [r0], #0x04 /* 0x40 */ 215 ldr r3, [r0], #0x04 /* 0x44 */ 216 pld [r0, #0x18] /* Prefetch 0x60 */ 217 strd r4, [r1], #0x08 218 ldr r4, [r0], #0x04 /* 0x48 */ 219 ldr r5, [r0], #0x04 /* 0x4c */ 220 strd r2, [r1], #0x08 221 ldr r2, [r0], #0x04 /* 0x50 */ 222 ldr r3, [r0], #0x04 /* 0x54 */ 223 strd r4, [r1], #0x08 224 ldr r4, [r0], #0x04 /* 0x58 */ 225 ldr r5, [r0], #0x04 /* 0x5c */ 226 strd r2, [r1], #0x08 227 ldr r2, [r0], #0x04 /* 0x60 */ 228 ldr r3, [r0], #0x04 /* 0x64 */ 229 pld [r0, #0x18] /* Prefetch 0x80 */ 230 strd r4, [r1], #0x08 231 ldr r4, [r0], #0x04 /* 0x68 */ 232 ldr r5, [r0], #0x04 /* 0x6c */ 233 strd r2, [r1], #0x08 234 ldr r2, [r0], #0x04 /* 0x70 */ 235 ldr r3, [r0], #0x04 /* 0x74 */ 236 strd r4, [r1], #0x08 237 ldr r4, [r0], #0x04 /* 0x78 */ 238 ldr r5, [r0], #0x04 /* 0x7c */ 239 strd r2, [r1], #0x08 240 subs ip, ip, #0x01 241 ldrgt r2, [r0], #0x04 /* 0x80 */ 242 ldrgt r3, [r0], #0x04 /* 0x84 */ 243 strd r4, [r1], #0x08 244 bgt 1b 245 ldmfd sp!, {r4, r5} 246 mov pc, lr 247 248 /* 249 * XSCALE version of bzero_page 250 */ 251 ENTRY(bzero_page) 252 mov r1, #PAGE_SIZE 253 mov r2, #0 254 mov r3, #0 255 1: strd r2, [r0], #8 /* 32 */ 256 strd r2, [r0], #8 257 strd r2, [r0], #8 258 strd r2, [r0], #8 259 strd r2, [r0], #8 /* 64 */ 260 strd r2, [r0], #8 261 strd r2, [r0], #8 262 strd r2, [r0], #8 263 strd r2, [r0], #8 /* 96 */ 264 strd r2, [r0], #8 265 strd r2, [r0], #8 266 strd r2, [r0], #8 267 strd r2, [r0], #8 /* 128 */ 268 strd r2, [r0], #8 269 strd r2, [r0], #8 270 strd r2, [r0], #8 271 subs r1, r1, #128 272 bne 1b 273 mov pc, lr 274 #endif /* __XSCALE__ */ 275
Indexes created Mon Oct 20 20:10:13 GMT 2025