| Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 1 | /* |
| 2 | * arch/ppc64/lib/memcpy.S |
| 3 | * |
| 4 | * Copyright (C) 2002 Paul Mackerras, IBM Corp. |
| 5 | * |
| 6 | * This program is free software; you can redistribute it and/or |
| 7 | * modify it under the terms of the GNU General Public License |
| 8 | * as published by the Free Software Foundation; either version |
| 9 | * 2 of the License, or (at your option) any later version. |
| 10 | */ |
| 11 | #include <asm/processor.h> |
| 12 | #include <asm/ppc_asm.h> |
| 13 | |
| 14 | .align 7 |
| 15 | _GLOBAL(memcpy) |
| 16 | mtcrf 0x01,r5 |
| 17 | cmpldi cr1,r5,16 |
| 18 | neg r6,r3 # LS 3 bits = # bytes to 8-byte dest bdry |
| 19 | andi. r6,r6,7 |
| 20 | dcbt 0,r4 |
| 21 | blt cr1,.Lshort_copy |
| 22 | bne .Ldst_unaligned |
| 23 | .Ldst_aligned: |
| 24 | andi. r0,r4,7 |
| 25 | addi r3,r3,-16 |
| 26 | bne .Lsrc_unaligned |
| 27 | srdi r7,r5,4 |
| 28 | ld r9,0(r4) |
| 29 | addi r4,r4,-8 |
| 30 | mtctr r7 |
| 31 | andi. r5,r5,7 |
| 32 | bf cr7*4+0,2f |
| 33 | addi r3,r3,8 |
| 34 | addi r4,r4,8 |
| 35 | mr r8,r9 |
| 36 | blt cr1,3f |
| 37 | 1: ld r9,8(r4) |
| 38 | std r8,8(r3) |
| 39 | 2: ldu r8,16(r4) |
| 40 | stdu r9,16(r3) |
| 41 | bdnz 1b |
| 42 | 3: std r8,8(r3) |
| 43 | beqlr |
| 44 | addi r3,r3,16 |
| 45 | ld r9,8(r4) |
| 46 | .Ldo_tail: |
| 47 | bf cr7*4+1,1f |
| 48 | rotldi r9,r9,32 |
| 49 | stw r9,0(r3) |
| 50 | addi r3,r3,4 |
| 51 | 1: bf cr7*4+2,2f |
| 52 | rotldi r9,r9,16 |
| 53 | sth r9,0(r3) |
| 54 | addi r3,r3,2 |
| 55 | 2: bf cr7*4+3,3f |
| 56 | rotldi r9,r9,8 |
| 57 | stb r9,0(r3) |
| 58 | 3: blr |
| 59 | |
| 60 | .Lsrc_unaligned: |
| 61 | srdi r6,r5,3 |
| 62 | addi r5,r5,-16 |
| 63 | subf r4,r0,r4 |
| 64 | srdi r7,r5,4 |
| 65 | sldi r10,r0,3 |
| 66 | cmpdi cr6,r6,3 |
| 67 | andi. r5,r5,7 |
| 68 | mtctr r7 |
| 69 | subfic r11,r10,64 |
| 70 | add r5,r5,r0 |
| 71 | |
| 72 | bt cr7*4+0,0f |
| 73 | |
| 74 | ld r9,0(r4) # 3+2n loads, 2+2n stores |
| 75 | ld r0,8(r4) |
| 76 | sld r6,r9,r10 |
| 77 | ldu r9,16(r4) |
| 78 | srd r7,r0,r11 |
| 79 | sld r8,r0,r10 |
| 80 | or r7,r7,r6 |
| 81 | blt cr6,4f |
| 82 | ld r0,8(r4) |
| 83 | # s1<< in r8, d0=(s0<<|s1>>) in r7, s3 in r0, s2 in r9, nix in r6 & r12 |
| 84 | b 2f |
| 85 | |
| 86 | 0: ld r0,0(r4) # 4+2n loads, 3+2n stores |
| 87 | ldu r9,8(r4) |
| 88 | sld r8,r0,r10 |
| 89 | addi r3,r3,-8 |
| 90 | blt cr6,5f |
| 91 | ld r0,8(r4) |
| 92 | srd r12,r9,r11 |
| 93 | sld r6,r9,r10 |
| 94 | ldu r9,16(r4) |
| 95 | or r12,r8,r12 |
| 96 | srd r7,r0,r11 |
| 97 | sld r8,r0,r10 |
| 98 | addi r3,r3,16 |
| 99 | beq cr6,3f |
| 100 | |
| 101 | # d0=(s0<<|s1>>) in r12, s1<< in r6, s2>> in r7, s2<< in r8, s3 in r9 |
| 102 | 1: or r7,r7,r6 |
| 103 | ld r0,8(r4) |
| 104 | std r12,8(r3) |
| 105 | 2: srd r12,r9,r11 |
| 106 | sld r6,r9,r10 |
| 107 | ldu r9,16(r4) |
| 108 | or r12,r8,r12 |
| 109 | stdu r7,16(r3) |
| 110 | srd r7,r0,r11 |
| 111 | sld r8,r0,r10 |
| 112 | bdnz 1b |
| 113 | |
| 114 | 3: std r12,8(r3) |
| 115 | or r7,r7,r6 |
| 116 | 4: std r7,16(r3) |
| 117 | 5: srd r12,r9,r11 |
| 118 | or r12,r8,r12 |
| 119 | std r12,24(r3) |
| 120 | beqlr |
| 121 | cmpwi cr1,r5,8 |
| 122 | addi r3,r3,32 |
| 123 | sld r9,r9,r10 |
| 124 | ble cr1,.Ldo_tail |
| 125 | ld r0,8(r4) |
| 126 | srd r7,r0,r11 |
| 127 | or r9,r7,r9 |
| 128 | b .Ldo_tail |
| 129 | |
| 130 | .Ldst_unaligned: |
| 131 | mtcrf 0x01,r6 # put #bytes to 8B bdry into cr7 |
| 132 | subf r5,r6,r5 |
| 133 | li r7,0 |
| 134 | cmpldi r1,r5,16 |
| 135 | bf cr7*4+3,1f |
| 136 | lbz r0,0(r4) |
| 137 | stb r0,0(r3) |
| 138 | addi r7,r7,1 |
| 139 | 1: bf cr7*4+2,2f |
| 140 | lhzx r0,r7,r4 |
| 141 | sthx r0,r7,r3 |
| 142 | addi r7,r7,2 |
| 143 | 2: bf cr7*4+1,3f |
| 144 | lwzx r0,r7,r4 |
| 145 | stwx r0,r7,r3 |
| 146 | 3: mtcrf 0x01,r5 |
| 147 | add r4,r6,r4 |
| 148 | add r3,r6,r3 |
| 149 | b .Ldst_aligned |
| 150 | |
| 151 | .Lshort_copy: |
| 152 | bf cr7*4+0,1f |
| 153 | lwz r0,0(r4) |
| 154 | lwz r9,4(r4) |
| 155 | addi r4,r4,8 |
| 156 | stw r0,0(r3) |
| 157 | stw r9,4(r3) |
| 158 | addi r3,r3,8 |
| 159 | 1: bf cr7*4+1,2f |
| 160 | lwz r0,0(r4) |
| 161 | addi r4,r4,4 |
| 162 | stw r0,0(r3) |
| 163 | addi r3,r3,4 |
| 164 | 2: bf cr7*4+2,3f |
| 165 | lhz r0,0(r4) |
| 166 | addi r4,r4,2 |
| 167 | sth r0,0(r3) |
| 168 | addi r3,r3,2 |
| 169 | 3: bf cr7*4+3,4f |
| 170 | lbz r0,0(r4) |
| 171 | stb r0,0(r3) |
| 172 | 4: blr |