| lh | 9ed821d | 2023-04-07 01:36:19 -0700 | [diff] [blame] | 1 | /* |
| 2 | * Copyright (C) 2004-2007 Atmel Corporation |
| 3 | * |
| 4 | * This file is subject to the terms and conditions of the GNU Lesser General |
| 5 | * Public License. See the file "COPYING.LIB" in the main directory of this |
| 6 | * archive for more details. |
| 7 | */ |
| 8 | |
| 9 | #define dst r12 |
| 10 | #define src r11 |
| 11 | #define len r10 |
| 12 | |
| 13 | .text |
| 14 | .global memmove |
| 15 | .type memmove, @function |
| 16 | memmove: |
| 17 | cp.w src, dst |
| 18 | brge HIDDEN_JUMPTARGET(memcpy) |
| 19 | |
| 20 | add dst, len |
| 21 | add src, len |
| 22 | pref src[-1] |
| 23 | |
| 24 | /* |
| 25 | * The rest is basically the same as in memcpy.S except that |
| 26 | * the direction is reversed. |
| 27 | */ |
| 28 | cp.w len, 32 |
| 29 | brge .Lmore_than_31 |
| 30 | |
| 31 | sub len, 1 |
| 32 | retlt r12 |
| 33 | 1: ld.ub r8, --src |
| 34 | st.b --dst, r8 |
| 35 | sub len, 1 |
| 36 | brge 1b |
| 37 | retal r12 |
| 38 | |
| 39 | .Lmore_than_31: |
| 40 | pushm r0-r7, lr |
| 41 | |
| 42 | /* Check alignment */ |
| 43 | mov r8, src |
| 44 | andl r8, 31, COH |
| 45 | brne .Lunaligned_src |
| 46 | mov r8, r12 |
| 47 | andl r8, 3, COH |
| 48 | brne .Lunaligned_dst |
| 49 | |
| 50 | .Laligned_copy: |
| 51 | sub len, 32 |
| 52 | brlt .Lless_than_32 |
| 53 | |
| 54 | 1: /* Copy 32 bytes at a time */ |
| 55 | sub src, 32 |
| 56 | ldm src, r0-r7 |
| 57 | sub dst, 32 |
| 58 | sub len, 32 |
| 59 | stm dst, r0-r7 |
| 60 | brge 1b |
| 61 | |
| 62 | .Lless_than_32: |
| 63 | /* Copy 16 more bytes if possible */ |
| 64 | sub len, -16 |
| 65 | brlt .Lless_than_16 |
| 66 | sub src, 16 |
| 67 | ldm src, r0-r3 |
| 68 | sub dst, 16 |
| 69 | sub len, 16 |
| 70 | stm dst, r0-r3 |
| 71 | |
| 72 | .Lless_than_16: |
| 73 | /* Do the remaining as byte copies */ |
| 74 | sub len, -16 |
| 75 | breq 2f |
| 76 | 1: ld.ub r0, --src |
| 77 | st.b --dst, r0 |
| 78 | sub len, 1 |
| 79 | brne 1b |
| 80 | |
| 81 | 2: popm r0-r7, pc |
| 82 | |
| 83 | .Lunaligned_src: |
| 84 | /* Make src cacheline-aligned. r8 = (src & 31) */ |
| 85 | sub len, r8 |
| 86 | 1: ld.ub r0, --src |
| 87 | st.b --dst, r0 |
| 88 | sub r8, 1 |
| 89 | brne 1b |
| 90 | |
| 91 | /* If dst is word-aligned, we're ready to go */ |
| 92 | pref src[-4] |
| 93 | mov r8, 3 |
| 94 | tst dst, r8 |
| 95 | breq .Laligned_copy |
| 96 | |
| 97 | .Lunaligned_dst: |
| 98 | /* src is aligned, but dst is not. Expect bad performance */ |
| 99 | sub len, 4 |
| 100 | brlt 2f |
| 101 | 1: ld.w r0, --src |
| 102 | st.w --dst, r0 |
| 103 | sub len, 4 |
| 104 | brge 1b |
| 105 | |
| 106 | 2: neg len |
| 107 | add pc, pc, len << 2 |
| 108 | .rept 3 |
| 109 | ld.ub r0, --src |
| 110 | st.b --dst, r0 |
| 111 | .endr |
| 112 | |
| 113 | popm r0-r7, pc |
| 114 | .size memmove, . - memmove |
| 115 | |
| 116 | libc_hidden_def(memmove) |