1 /* Copyright (c) 2012-2013, Linaro Limited 2 All rights reserved. 3 4 Redistribution and use in source and binary forms, with or without 5 modification, are permitted provided that the following conditions are met: 6 * Redistributions of source code must retain the above copyright 7 notice, this list of conditions and the following disclaimer. 8 * Redistributions in binary form must reproduce the above copyright 9 notice, this list of conditions and the following disclaimer in the 10 documentation and/or other materials provided with the distribution. 11 * Neither the name of the Linaro nor the 12 names of its contributors may be used to endorse or promote products 13 derived from this software without specific prior written permission. 14 15 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS 16 "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT 17 LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR 18 A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT 19 HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, 20 SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT 21 LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 22 DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 23 THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 24 (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE 25 OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ 26 27 /* 28 * Copyright (c) 2015 ARM Ltd 29 * All rights reserved. 30 * 31 * Redistribution and use in source and binary forms, with or without 32 * modification, are permitted provided that the following conditions 33 * are met: 34 * 1. Redistributions of source code must retain the above copyright 35 * notice, this list of conditions and the following disclaimer. 36 * 2. Redistributions in binary form must reproduce the above copyright 37 * notice, this list of conditions and the following disclaimer in the 38 * documentation and/or other materials provided with the distribution. 39 * 3. The name of the company may not be used to endorse or promote 40 * products derived from this software without specific prior written 41 * permission. 42 * 43 * THIS SOFTWARE IS PROVIDED BY ARM LTD ``AS IS'' AND ANY EXPRESS OR IMPLIED 44 * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF 45 * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. 46 * IN NO EVENT SHALL ARM LTD BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, 47 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED 48 * TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR 49 * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF 50 * LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING 51 * NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS 52 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 53 */ 54 55 /* Assumptions: 56 * 57 * ARMv8-a, AArch64, unaligned accesses. 58 * 59 */ 60 61 #include <private/bionic_asm.h> 62 63 #define dstin x0 64 #define src x1 65 #define count x2 66 #define dst x3 67 #define srcend x4 68 #define dstend x5 69 #define A_l x6 70 #define A_lw w6 71 #define A_h x7 72 #define A_hw w7 73 #define B_l x8 74 #define B_lw w8 75 #define B_h x9 76 #define C_l x10 77 #define C_h x11 78 #define D_l x12 79 #define D_h x13 80 #define E_l src 81 #define E_h count 82 #define F_l srcend 83 #define F_h dst 84 #define tmp1 x9 85 86 #define L(l) .L ## l 87 88 /* Copies are split into 3 main cases: small copies of up to 16 bytes, 89 medium copies of 17..96 bytes which are fully unrolled. Large copies 90 of more than 96 bytes align the destination and use an unrolled loop 91 processing 64 bytes per iteration. 92 Small and medium copies read all data before writing, allowing any 93 kind of overlap, and memmove tailcalls memcpy for these cases as 94 well as non-overlapping copies. 95 */ 96 97 prfm PLDL1KEEP, [src] 98 add srcend, src, count 99 add dstend, dstin, count 100 cmp count, 16 101 b.ls L(copy16) 102 cmp count, 96 103 b.hi L(copy_long) 104 105 /* Medium copies: 17..96 bytes. */ 106 sub tmp1, count, 1 107 ldp A_l, A_h, [src] 108 tbnz tmp1, 6, L(copy96) 109 ldp D_l, D_h, [srcend, -16] 110 tbz tmp1, 5, 1f 111 ldp B_l, B_h, [src, 16] 112 ldp C_l, C_h, [srcend, -32] 113 stp B_l, B_h, [dstin, 16] 114 stp C_l, C_h, [dstend, -32] 115 1: 116 stp A_l, A_h, [dstin] 117 stp D_l, D_h, [dstend, -16] 118 ret 119 120 .p2align 4 121 122 /* Small copies: 0..16 bytes. */ 123 L(copy16): 124 cmp count, 8 125 b.lo 1f 126 ldr A_l, [src] 127 ldr A_h, [srcend, -8] 128 str A_l, [dstin] 129 str A_h, [dstend, -8] 130 ret 131 .p2align 4 132 1: 133 tbz count, 2, 1f 134 ldr A_lw, [src] 135 ldr A_hw, [srcend, -4] 136 str A_lw, [dstin] 137 str A_hw, [dstend, -4] 138 ret 139 140 /* Copy 0..3 bytes. Use a branchless sequence that copies the same 141 byte 3 times if count==1, or the 2nd byte twice if count==2. */ 142 1: 143 cbz count, 2f 144 lsr tmp1, count, 1 145 ldrb A_lw, [src] 146 ldrb A_hw, [srcend, -1] 147 ldrb B_lw, [src, tmp1] 148 strb A_lw, [dstin] 149 strb B_lw, [dstin, tmp1] 150 strb A_hw, [dstend, -1] 151 2: ret 152 153 .p2align 4 154 /* Copy 64..96 bytes. Copy 64 bytes from the start and 155 32 bytes from the end. */ 156 L(copy96): 157 ldp B_l, B_h, [src, 16] 158 ldp C_l, C_h, [src, 32] 159 ldp D_l, D_h, [src, 48] 160 ldp E_l, E_h, [srcend, -32] 161 ldp F_l, F_h, [srcend, -16] 162 stp A_l, A_h, [dstin] 163 stp B_l, B_h, [dstin, 16] 164 stp C_l, C_h, [dstin, 32] 165 stp D_l, D_h, [dstin, 48] 166 stp E_l, E_h, [dstend, -32] 167 stp F_l, F_h, [dstend, -16] 168 ret 169 170 /* Align DST to 16 byte alignment so that we don't cross cache line 171 boundaries on both loads and stores. There are at least 96 bytes 172 to copy, so copy 16 bytes unaligned and then align. The loop 173 copies 64 bytes per iteration and prefetches one iteration ahead. */ 174 175 .p2align 4 176 L(copy_long): 177 and tmp1, dstin, 15 178 bic dst, dstin, 15 179 ldp D_l, D_h, [src] 180 sub src, src, tmp1 181 add count, count, tmp1 /* Count is now 16 too large. */ 182 ldp A_l, A_h, [src, 16] 183 stp D_l, D_h, [dstin] 184 ldp B_l, B_h, [src, 32] 185 ldp C_l, C_h, [src, 48] 186 ldp D_l, D_h, [src, 64]! 187 subs count, count, 128 + 16 /* Test and readjust count. */ 188 b.ls 2f 189 1: 190 stp A_l, A_h, [dst, 16] 191 ldp A_l, A_h, [src, 16] 192 stp B_l, B_h, [dst, 32] 193 ldp B_l, B_h, [src, 32] 194 stp C_l, C_h, [dst, 48] 195 ldp C_l, C_h, [src, 48] 196 stp D_l, D_h, [dst, 64]! 197 ldp D_l, D_h, [src, 64]! 198 subs count, count, 64 199 b.hi 1b 200 201 /* Write the last full set of 64 bytes. The remainder is at most 64 202 bytes, so it is safe to always copy 64 bytes from the end even if 203 there is just 1 byte left. */ 204 2: 205 ldp E_l, E_h, [srcend, -64] 206 stp A_l, A_h, [dst, 16] 207 ldp A_l, A_h, [srcend, -48] 208 stp B_l, B_h, [dst, 32] 209 ldp B_l, B_h, [srcend, -32] 210 stp C_l, C_h, [dst, 48] 211 ldp C_l, C_h, [srcend, -16] 212 stp D_l, D_h, [dst, 64] 213 stp E_l, E_h, [dstend, -64] 214 stp A_l, A_h, [dstend, -48] 215 stp B_l, B_h, [dstend, -32] 216 stp C_l, C_h, [dstend, -16] 217 ret 218
