1 // Inferno's libkern/memmove-386.s 2 // http://code.google.com/p/inferno-os/source/browse/libkern/memmove-386.s 3 // 4 // Copyright 1994-1999 Lucent Technologies Inc. All rights reserved. 5 // Revisions Copyright 2000-2007 Vita Nuova Holdings Limited (www.vitanuova.com). All rights reserved. 6 // Portions Copyright 2009 The Go Authors. All rights reserved. 7 // 8 // Permission is hereby granted, free of charge, to any person obtaining a copy 9 // of this software and associated documentation files (the "Software"), to deal 10 // in the Software without restriction, including without limitation the rights 11 // to use, copy, modify, merge, publish, distribute, sublicense, and/or sell 12 // copies of the Software, and to permit persons to whom the Software is 13 // furnished to do so, subject to the following conditions: 14 // 15 // The above copyright notice and this permission notice shall be included in 16 // all copies or substantial portions of the Software. 17 // 18 // THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR 19 // IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 20 // FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE 21 // AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER 22 // LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, 23 // OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN 24 // THE SOFTWARE. 25 26 #include "textflag.h" 27 28 TEXT runtimememmove(SB), NOSPLIT, $0-12 29 MOVL to+0(FP), DI 30 MOVL from+4(FP), SI 31 MOVL n+8(FP), BX 32 33 // REP instructions have a high startup cost, so we handle small sizes 34 // with some straightline code. The REP MOVSL instruction is really fast 35 // for large sizes. The cutover is approximately 1K. 36 tail: 37 TESTL BX, BX 38 JEQ move_0 39 CMPL BX, $2 40 JBE move_1or2 41 CMPL BX, $4 42 JBE move_3or4 43 CMPL BX, $8 44 JBE move_5through8 45 CMPL BX, $16 46 JBE move_9through16 47 48 /* 49 * check and set for backwards 50 */ 51 CMPL SI, DI 52 JLS back 53 54 /* 55 * forward copy loop 56 */ 57 forward: 58 MOVL BX, CX 59 SHRL $2, CX 60 ANDL $3, BX 61 62 REP; MOVSL 63 JMP tail 64 /* 65 * check overlap 66 */ 67 back: 68 MOVL SI, CX 69 ADDL BX, CX 70 CMPL CX, DI 71 JLS forward 72 /* 73 * whole thing backwards has 74 * adjusted addresses 75 */ 76 77 ADDL BX, DI 78 ADDL BX, SI 79 STD 80 81 /* 82 * copy 83 */ 84 MOVL BX, CX 85 SHRL $2, CX 86 ANDL $3, BX 87 88 SUBL $4, DI 89 SUBL $4, SI 90 REP; MOVSL 91 92 CLD 93 ADDL $4, DI 94 ADDL $4, SI 95 SUBL BX, DI 96 SUBL BX, SI 97 JMP tail 98 99 move_1or2: 100 MOVB (SI), AX 101 MOVB -1(SI)(BX*1), CX 102 MOVB AX, (DI) 103 MOVB CX, -1(DI)(BX*1) 104 RET 105 move_0: 106 RET 107 move_3or4: 108 MOVW (SI), AX 109 MOVW -2(SI)(BX*1), CX 110 MOVW AX, (DI) 111 MOVW CX, -2(DI)(BX*1) 112 RET 113 move_5through8: 114 MOVL (SI), AX 115 MOVL -4(SI)(BX*1), CX 116 MOVL AX, (DI) 117 MOVL CX, -4(DI)(BX*1) 118 RET 119 move_9through16: 120 MOVL (SI), AX 121 MOVL 4(SI), CX 122 MOVL -8(SI)(BX*1), DX 123 MOVL -4(SI)(BX*1), BP 124 MOVL AX, (DI) 125 MOVL CX, 4(DI) 126 MOVL DX, -8(DI)(BX*1) 127 MOVL BP, -4(DI)(BX*1) 128 RET 129
