1/* 2 * CDDL HEADER START 3 * 4 * The contents of this file are subject to the terms of the 5 * Common Development and Distribution License, Version 1.0 only 6 * (the "License"). You may not use this file except in compliance 7 * with the License. 8 * 9 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE 10 * or http://www.opensolaris.org/os/licensing. 11 * See the License for the specific language governing permissions 12 * and limitations under the License. 13 * 14 * When distributing Covered Code, include this CDDL HEADER in each 15 * file and include the License file at usr/src/OPENSOLARIS.LICENSE. 16 * If applicable, add the following below this CDDL HEADER, with the 17 * fields enclosed by brackets "[]" replaced with your own identifying 18 * information: Portions Copyright [yyyy] [name of copyright owner] 19 * 20 * CDDL HEADER END 21 */ 22/* 23 * Copyright 2004 Sun Microsystems, Inc. All rights reserved. 24 * Use is subject to license terms. 25 */ 26 27 .ident "%Z%%M% %I% %E% SMI" 28 29 .file "%M%" 30 31/ 32/ strncat(s1, s2, n) 33/ 34/ Concatenates s2 on the end of s1. s1's space must be large enough. 35/ At most n characters are moved. 36/ Returns s1. 37/ 38/ Fast assembly language version of the following C-program strncat 39/ which represents the `standard' for the C-library. 40/ 41/ char * 42/ strncat(char *s1, const char *s2, size_t n) 43/ { 44/ char *os1 = s1; 45/ 46/ n++; 47/ while (*s1++) 48/ ; 49/ --s1; 50/ while (*s1++ = *s2++) 51/ if (--n == 0) { 52/ s1[-1] = '\0'; 53/ break; 54/ } 55/ return (os1); 56/ } 57/ 58/ In this assembly language version, the following expression is used 59/ to check if a 32-bit word data contains a null byte or not: 60/ (((A & 0x7f7f7f7f) + 0x7f7f7f7f) | A) & 0x80808080 61/ If the above expression geneates a value other than 0x80808080, 62/ that means the 32-bit word data contains a null byte. 63/ 64/ The above has been extended for 64-bit support. 65/ 66 67#include "SYS.h" 68 69 ENTRY(strncat) /* (char *, char *, size_t) */ 70 movq %rdi, %rax / save return value 71 movabsq $0x7f7f7f7f7f7f7f7f, %r8 / %r8 = 0x7f... 72 movq %r8, %r9 73 notq %r9 / %r9 = 0x80... 74 testq $7, %rdi / if %rdi not quadword aligned 75 jnz .L1 / goto .L1 76 .align 4 77.L2: 78 movq (%rdi), %r11 / move 1 quadword from (%rdi) to %r11 79 movq %r8, %rcx 80 andq %r11, %rcx / %rcx = %r11 & 0x7f7f7f7f 81 addq $8, %rdi / next quadword 82 addq %r8, %rcx / %rcx += 0x7f7f7f7f 83 orq %r11, %rcx / %rcx |= %r11 84 andq %r9, %rcx / %rcx &= 0x80808080 85 cmpq %r9, %rcx / if no null byte in this quadword 86 je .L2 / goto .L2 87 subq $8, %rdi / post-incremented 88.L1: 89 cmpb $0, (%rdi) / if a byte in (%rdi) is null 90 je .L3 / goto .L3 91 incq %rdi / next byte 92 testq $7, %rdi / if %rdi not quadword aligned 93 jnz .L1 / goto .L1 94 jmp .L2 / goto .L2 (%rdi quadword aligned) 95 .align 4 96.L3: 97 / %rdi points to a null byte in destination string 98 99 testq $7, %rsi / if %rsi not quadword aligned 100 jnz .L4 / goto .L4 101 cmpq $8, %rdx / if number of bytes < 8 102 jb .L7 / goto .L7 103 .align 4 104.L5: 105 movq (%rsi), %r11 / move 1 quadword from (%rsi) to %r11 106 movq %r8, %rcx 107 andq %r11, %rcx / %rcx = %r11 & 0x7f7f7f7f 108 addq $8, %rsi / next quadword 109 addq %r8, %rcx / %rcx += 0x7f7f7f7f 110 orq %r11, %rcx / %rcx |= %r11 111 andq %r9, %rcx / %rcx &= 0x80808080 112 cmpq %r9, %rcx / if null byte in this quadword 113 jne .L6 / goto .L6 114 movq %r11, (%rdi) / copy this quadword to (%rdi) 115 subq $8, %rdx / decrement number of bytes by 8 116 addq $8, %rdi / next quadword 117 cmpq $8, %rdx / if number of bytes >= 8 118 jae .L5 / goto .L5 119 jmp .L7 / goto .L7 120.L6: 121 subq $8, %rsi / post-incremented 122 .align 4 123.L7: 124 / number of bytes < 8 or a null byte found in the quadword 125 cmpq $0, %rdx / if number of bytes == 0 126 jz .L8 / goto .L8 (finished) 127 movb (%rsi), %r11b / %r11b = a byte in (%rsi) 128 decq %rdx / decrement number of bytes by 1 129 movb %r11b, (%rdi) / copy %r11b to (%rdi) 130 incq %rsi / next byte 131 incq %rdi / next byte 132 cmpb $0, %r11b / compare %r11b with a null byte 133 je .L9 / if %r11b is a null, goto .L9 134 jmp .L7 / goto .L7 135 .align 4 136 137.L4: 138 / %rsi not aligned 139 cmpq $0, %rdx / if number of bytes == 0 140 jz .L8 / goto .L8 (finished) 141 movb (%rsi), %r11b / %r11b = a byte in (%rsi) 142 decq %rdx / decrement number of bytes by 1 143 movb %r11b, (%rdi) / copy %r11b to (%rdi) 144 incq %rdi / next byte 145 incq %rsi / next byte 146 cmpb $0, %r11b / compare %r11b with a null byte 147 je .L9 / if %r11b is a null, goto .L9 148 jmp .L4 / goto .L4 149 .align 4 150.L8: 151 movb $0, (%rdi) / null termination 152.L9: 153 ret 154 SET_SIZE(strncat) 155