1 /*- 2 * Copyright (c) 2003-2007 Tim Kientzle 3 * All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 14 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR 15 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES 16 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. 17 * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, 18 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT 19 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 20 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 21 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 22 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF 23 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 24 */ 25 26 #include "archive_platform.h" 27 __FBSDID("$FreeBSD: src/lib/libarchive/archive_read_open_filename.c,v 1.21 2008/02/19 06:10:48 kientzle Exp $"); 28 29 #ifdef HAVE_SYS_STAT_H 30 #include <sys/stat.h> 31 #endif 32 #ifdef HAVE_ERRNO_H 33 #include <errno.h> 34 #endif 35 #ifdef HAVE_FCNTL_H 36 #include <fcntl.h> 37 #endif 38 #ifdef HAVE_STDLIB_H 39 #include <stdlib.h> 40 #endif 41 #ifdef HAVE_STRING_H 42 #include <string.h> 43 #endif 44 #ifdef HAVE_UNISTD_H 45 #include <unistd.h> 46 #endif 47 48 #include "archive.h" 49 50 #ifndef O_BINARY 51 #define O_BINARY 0 52 #endif 53 54 struct read_file_data { 55 int fd; 56 size_t block_size; 57 void *buffer; 58 mode_t st_mode; /* Mode bits for opened file. */ 59 char can_skip; /* This file supports skipping. */ 60 char filename[1]; /* Must be last! */ 61 }; 62 63 static int file_close(struct archive *, void *); 64 static ssize_t file_read(struct archive *, void *, const void **buff); 65 #if ARCHIVE_API_VERSION < 2 66 static ssize_t file_skip(struct archive *, void *, size_t request); 67 #else 68 static off_t file_skip(struct archive *, void *, off_t request); 69 #endif 70 71 int 72 archive_read_open_file(struct archive *a, const char *filename, 73 size_t block_size) 74 { 75 return (archive_read_open_filename(a, filename, block_size)); 76 } 77 78 int 79 archive_read_open_filename(struct archive *a, const char *filename, 80 size_t block_size) 81 { 82 struct stat st; 83 struct read_file_data *mine; 84 void *b; 85 int fd; 86 87 if (filename == NULL || filename[0] == '\0') 88 return (archive_read_open_fd(a, 0, block_size)); 89 90 fd = open(filename, O_RDONLY | O_BINARY); 91 if (fd < 0) { 92 archive_set_error(a, errno, "Failed to open '%s'", filename); 93 return (ARCHIVE_FATAL); 94 } 95 if (fstat(fd, &st) != 0) { 96 archive_set_error(a, errno, "Can't stat '%s'", filename); 97 return (ARCHIVE_FATAL); 98 } 99 100 mine = (struct read_file_data *)malloc(sizeof(*mine) + strlen(filename)); 101 b = malloc(block_size); 102 if (mine == NULL || b == NULL) { 103 archive_set_error(a, ENOMEM, "No memory"); 104 free(mine); 105 free(b); 106 return (ARCHIVE_FATAL); 107 } 108 strcpy(mine->filename, filename); 109 mine->block_size = block_size; 110 mine->buffer = b; 111 mine->fd = fd; 112 /* Remember mode so close can decide whether to flush. */ 113 mine->st_mode = st.st_mode; 114 /* If we're reading a file from disk, ensure that we don't 115 overwrite it with an extracted file. */ 116 if (S_ISREG(st.st_mode)) { 117 archive_read_extract_set_skip_file(a, st.st_dev, st.st_ino); 118 /* 119 * Skip is a performance optimization for anything 120 * that supports lseek(). Generally, that only 121 * includes regular files and possibly raw disk 122 * devices, but there's no good portable way to detect 123 * raw disks. 124 */ 125 mine->can_skip = 1; 126 } else 127 mine->can_skip = 0; 128 return (archive_read_open2(a, mine, 129 NULL, file_read, file_skip, file_close)); 130 } 131 132 static ssize_t 133 file_read(struct archive *a, void *client_data, const void **buff) 134 { 135 struct read_file_data *mine = (struct read_file_data *)client_data; 136 ssize_t bytes_read; 137 138 *buff = mine->buffer; 139 bytes_read = read(mine->fd, mine->buffer, mine->block_size); 140 if (bytes_read < 0) { 141 archive_set_error(a, errno, "Error reading '%s'", 142 mine->filename); 143 } 144 return (bytes_read); 145 } 146 147 #if ARCHIVE_API_VERSION < 2 148 static ssize_t 149 file_skip(struct archive *a, void *client_data, size_t request) 150 #else 151 static off_t 152 file_skip(struct archive *a, void *client_data, off_t request) 153 #endif 154 { 155 struct read_file_data *mine = (struct read_file_data *)client_data; 156 off_t old_offset, new_offset; 157 158 if (!mine->can_skip) /* We can't skip, so ... */ 159 return (0); /* ... skip zero bytes. */ 160 161 /* Reduce request to the next smallest multiple of block_size */ 162 request = (request / mine->block_size) * mine->block_size; 163 if (request == 0) 164 return (0); 165 166 /* 167 * Hurray for lazy evaluation: if the first lseek fails, the second 168 * one will not be executed. 169 */ 170 if (((old_offset = lseek(mine->fd, 0, SEEK_CUR)) < 0) || 171 ((new_offset = lseek(mine->fd, request, SEEK_CUR)) < 0)) 172 { 173 /* If skip failed once, it will probably fail again. */ 174 mine->can_skip = 0; 175 176 if (errno == ESPIPE) 177 { 178 /* 179 * Failure to lseek() can be caused by the file 180 * descriptor pointing to a pipe, socket or FIFO. 181 * Return 0 here, so the compression layer will use 182 * read()s instead to advance the file descriptor. 183 * It's slower of course, but works as well. 184 */ 185 return (0); 186 } 187 /* 188 * There's been an error other than ESPIPE. This is most 189 * likely caused by a programmer error (too large request) 190 * or a corrupted archive file. 191 */ 192 archive_set_error(a, errno, "Error seeking in '%s'", 193 mine->filename); 194 return (-1); 195 } 196 return (new_offset - old_offset); 197 } 198 199 static int 200 file_close(struct archive *a, void *client_data) 201 { 202 struct read_file_data *mine = (struct read_file_data *)client_data; 203 204 (void)a; /* UNUSED */ 205 206 /* Only flush and close if open succeeded. */ 207 if (mine->fd >= 0) { 208 /* 209 * Sometimes, we should flush the input before closing. 210 * Regular files: faster to just close without flush. 211 * Devices: must not flush (user might need to 212 * read the "next" item on a non-rewind device). 213 * Pipes and sockets: must flush (otherwise, the 214 * program feeding the pipe or socket may complain). 215 * Here, I flush everything except for regular files and 216 * device nodes. 217 */ 218 if (!S_ISREG(mine->st_mode) 219 && !S_ISCHR(mine->st_mode) 220 && !S_ISBLK(mine->st_mode)) { 221 ssize_t bytesRead; 222 do { 223 bytesRead = read(mine->fd, mine->buffer, 224 mine->block_size); 225 } while (bytesRead > 0); 226 } 227 close(mine->fd); 228 } 229 free(mine->buffer); 230 free(mine); 231 return (ARCHIVE_OK); 232 } 233