| 1 | /* |
| 2 | * Copyright (c) 1998 Michael Smith. |
| 3 | * Copyright (c) 2000 Maxim Sobolev |
| 4 | * All rights reserved. |
| 5 | * |
| 6 | * Redistribution and use in source and binary forms, with or without |
| 7 | * modification, are permitted provided that the following conditions |
| 8 | * are met: |
| 9 | * 1. Redistributions of source code must retain the above copyright |
| 10 | * notice, this list of conditions and the following disclaimer. |
| 11 | * 2. Redistributions in binary form must reproduce the above copyright |
| 12 | * notice, this list of conditions and the following disclaimer in the |
| 13 | * documentation and/or other materials provided with the distribution. |
| 14 | * |
| 15 | * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND |
| 16 | * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
| 17 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE |
| 18 | * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE |
| 19 | * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
| 20 | * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS |
| 21 | * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) |
| 22 | * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT |
| 23 | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY |
| 24 | * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF |
| 25 | * SUCH DAMAGE. |
| 26 | * |
| 27 | * $FreeBSD: src/lib/libstand/bzipfs.c,v 1.2.2.3 2002/04/08 13:50:09 sobomax Exp $ |
| 28 | * $DragonFly: src/lib/libstand/bzipfs.c,v 1.4 2007/05/13 18:33:56 swildner Exp $ |
| 29 | * |
| 30 | */ |
| 31 | |
| 32 | #include "stand.h" |
| 33 | |
| 34 | #include <sys/stat.h> |
| 35 | #include <string.h> |
| 36 | #include <bzlib.h> |
| 37 | |
| 38 | #define BZ_BUFSIZE 2048 /* XXX larger? */ |
| 39 | |
| 40 | struct bz_file |
| 41 | { |
| 42 | int bzf_rawfd; |
| 43 | bz_stream bzf_bzstream; |
| 44 | char bzf_buf[BZ_BUFSIZE]; |
| 45 | }; |
| 46 | |
| 47 | static int bzf_fill(struct bz_file *z); |
| 48 | static int bzf_open(const char *path, struct open_file *f); |
| 49 | static int bzf_close(struct open_file *f); |
| 50 | static int bzf_read(struct open_file *f, void *buf, size_t size, size_t *resid); |
| 51 | static off_t bzf_seek(struct open_file *f, off_t offset, int where); |
| 52 | static int bzf_stat(struct open_file *f, struct stat *sb); |
| 53 | |
| 54 | struct fs_ops bzipfs_fsops = { |
| 55 | "bzip", |
| 56 | bzf_open, |
| 57 | bzf_close, |
| 58 | bzf_read, |
| 59 | null_write, |
| 60 | bzf_seek, |
| 61 | bzf_stat, |
| 62 | null_readdir |
| 63 | }; |
| 64 | |
| 65 | #if 0 |
| 66 | void * |
| 67 | calloc(int items, size_t size) |
| 68 | { |
| 69 | return(malloc(items * size)); |
| 70 | } |
| 71 | #endif |
| 72 | |
| 73 | static int |
| 74 | bzf_fill(struct bz_file *bzf) |
| 75 | { |
| 76 | int result; |
| 77 | int req; |
| 78 | |
| 79 | req = BZ_BUFSIZE - bzf->bzf_bzstream.avail_in; |
| 80 | result = 0; |
| 81 | |
| 82 | /* If we need more */ |
| 83 | if (req > 0) { |
| 84 | /* move old data to bottom of buffer */ |
| 85 | if (req < BZ_BUFSIZE) |
| 86 | bcopy(bzf->bzf_buf + req, bzf->bzf_buf, BZ_BUFSIZE - req); |
| 87 | |
| 88 | /* read to fill buffer and update availibility data */ |
| 89 | result = read(bzf->bzf_rawfd, bzf->bzf_buf + bzf->bzf_bzstream.avail_in, req); |
| 90 | bzf->bzf_bzstream.next_in = bzf->bzf_buf; |
| 91 | if (result >= 0) |
| 92 | bzf->bzf_bzstream.avail_in += result; |
| 93 | } |
| 94 | return(result); |
| 95 | } |
| 96 | |
| 97 | /* |
| 98 | * Adapted from get_byte/check_header in libz |
| 99 | * |
| 100 | * Returns 0 if the header is OK, nonzero if not. |
| 101 | */ |
| 102 | static int |
| 103 | get_byte(struct bz_file *bzf) |
| 104 | { |
| 105 | if ((bzf->bzf_bzstream.avail_in == 0) && (bzf_fill(bzf) == -1)) |
| 106 | return(-1); |
| 107 | bzf->bzf_bzstream.avail_in--; |
| 108 | return(*(bzf->bzf_bzstream.next_in)++); |
| 109 | } |
| 110 | |
| 111 | static int bz_magic[3] = {'B', 'Z', 'h'}; /* bzip2 magic header */ |
| 112 | |
| 113 | static int |
| 114 | check_header(struct bz_file *bzf) |
| 115 | { |
| 116 | unsigned int len; |
| 117 | int c; |
| 118 | |
| 119 | /* Check the bzip2 magic header */ |
| 120 | for (len = 0; len < 3; len++) { |
| 121 | c = get_byte(bzf); |
| 122 | if (c != bz_magic[len]) { |
| 123 | return(1); |
| 124 | } |
| 125 | } |
| 126 | /* Check that the block size is valid */ |
| 127 | c = get_byte(bzf); |
| 128 | if (c < '1' || c > '9') |
| 129 | return(1); |
| 130 | |
| 131 | /* Put back bytes that we've took from the input stream */ |
| 132 | bzf->bzf_bzstream.next_in -= 4; |
| 133 | bzf->bzf_bzstream.avail_in += 4; |
| 134 | |
| 135 | return(0); |
| 136 | } |
| 137 | |
| 138 | static int |
| 139 | bzf_open(const char *fname, struct open_file *f) |
| 140 | { |
| 141 | static char *bzfname; |
| 142 | int rawfd; |
| 143 | struct bz_file *bzf; |
| 144 | char *cp; |
| 145 | int error; |
| 146 | struct stat sb; |
| 147 | |
| 148 | /* Have to be in "just read it" mode */ |
| 149 | if (f->f_flags != F_READ) |
| 150 | return(EPERM); |
| 151 | |
| 152 | /* If the name already ends in .gz or .bz2, ignore it */ |
| 153 | if ((cp = strrchr(fname, '.')) && (!strcmp(cp, ".gz") |
| 154 | || !strcmp(cp, ".bz2") || !strcmp(cp, ".split"))) |
| 155 | return(ENOENT); |
| 156 | |
| 157 | /* Construct new name */ |
| 158 | bzfname = malloc(strlen(fname) + 5); |
| 159 | sprintf(bzfname, "%s.bz2", fname); |
| 160 | |
| 161 | /* Try to open the compressed datafile */ |
| 162 | rawfd = open(bzfname, O_RDONLY); |
| 163 | free(bzfname); |
| 164 | if (rawfd == -1) |
| 165 | return(ENOENT); |
| 166 | |
| 167 | if (fstat(rawfd, &sb) < 0) { |
| 168 | printf("bzf_open: stat failed\n"); |
| 169 | close(rawfd); |
| 170 | return(ENOENT); |
| 171 | } |
| 172 | if (!S_ISREG(sb.st_mode)) { |
| 173 | printf("bzf_open: not a file\n"); |
| 174 | close(rawfd); |
| 175 | return(EISDIR); /* best guess */ |
| 176 | } |
| 177 | |
| 178 | /* Allocate a bz_file structure, populate it */ |
| 179 | bzf = malloc(sizeof(struct bz_file)); |
| 180 | bzero(bzf, sizeof(struct bz_file)); |
| 181 | bzf->bzf_rawfd = rawfd; |
| 182 | |
| 183 | /* Verify that the file is bzipped (XXX why do this afterwards?) */ |
| 184 | if (check_header(bzf)) { |
| 185 | close(bzf->bzf_rawfd); |
| 186 | BZ2_bzDecompressEnd(&(bzf->bzf_bzstream)); |
| 187 | free(bzf); |
| 188 | return(EFTYPE); |
| 189 | } |
| 190 | |
| 191 | /* Initialise the inflation engine */ |
| 192 | if ((error = BZ2_bzDecompressInit(&(bzf->bzf_bzstream), 0, 1)) != BZ_OK) { |
| 193 | printf("bzf_open: BZ2_bzDecompressInit returned %d\n", error); |
| 194 | close(bzf->bzf_rawfd); |
| 195 | free(bzf); |
| 196 | return(EIO); |
| 197 | } |
| 198 | |
| 199 | /* Looks OK, we'll take it */ |
| 200 | f->f_fsdata = bzf; |
| 201 | return(0); |
| 202 | } |
| 203 | |
| 204 | static int |
| 205 | bzf_close(struct open_file *f) |
| 206 | { |
| 207 | struct bz_file *bzf = (struct bz_file *)f->f_fsdata; |
| 208 | |
| 209 | f->f_fsdata = NULL; |
| 210 | if (bzf) { |
| 211 | BZ2_bzDecompressEnd(&(bzf->bzf_bzstream)); |
| 212 | close(bzf->bzf_rawfd); |
| 213 | free(bzf); |
| 214 | } |
| 215 | return(0); |
| 216 | } |
| 217 | |
| 218 | static int |
| 219 | bzf_read(struct open_file *f, void *buf, size_t size, size_t *resid) |
| 220 | { |
| 221 | struct bz_file *bzf = (struct bz_file *)f->f_fsdata; |
| 222 | int error; |
| 223 | |
| 224 | bzf->bzf_bzstream.next_out = buf; /* where and how much */ |
| 225 | bzf->bzf_bzstream.avail_out = size; |
| 226 | |
| 227 | while (bzf->bzf_bzstream.avail_out) { |
| 228 | if ((bzf->bzf_bzstream.avail_in == 0) && (bzf_fill(bzf) == -1)) { |
| 229 | printf("bzf_read: fill error\n"); |
| 230 | return(-1); |
| 231 | } |
| 232 | if (bzf->bzf_bzstream.avail_in == 0) { /* oops, unexpected EOF */ |
| 233 | printf("bzf_read: unexpected EOF\n"); |
| 234 | break; |
| 235 | } |
| 236 | |
| 237 | error = BZ2_bzDecompress(&bzf->bzf_bzstream); /* decompression pass */ |
| 238 | if (error == BZ_STREAM_END) { /* EOF, all done */ |
| 239 | break; |
| 240 | } |
| 241 | if (error != BZ_OK) { /* argh, decompression error */ |
| 242 | printf("bzf_read: BZ2_bzDecompress returned %d\n", error); |
| 243 | errno = EIO; |
| 244 | return(-1); |
| 245 | } |
| 246 | } |
| 247 | if (resid != NULL) |
| 248 | *resid = bzf->bzf_bzstream.avail_out; |
| 249 | return(0); |
| 250 | } |
| 251 | |
| 252 | static off_t |
| 253 | bzf_seek(struct open_file *f, off_t offset, int where) |
| 254 | { |
| 255 | struct bz_file *bzf = (struct bz_file *)f->f_fsdata; |
| 256 | off_t target; |
| 257 | char discard[16]; |
| 258 | |
| 259 | switch (where) { |
| 260 | case SEEK_SET: |
| 261 | target = offset; |
| 262 | break; |
| 263 | case SEEK_CUR: |
| 264 | target = offset + bzf->bzf_bzstream.total_out_lo32; |
| 265 | break; |
| 266 | default: |
| 267 | target = -1; |
| 268 | } |
| 269 | |
| 270 | /* Can we get there from here? */ |
| 271 | if (target < bzf->bzf_bzstream.total_out_lo32) { |
| 272 | errno = EOFFSET; |
| 273 | return -1; |
| 274 | } |
| 275 | |
| 276 | /* skip forwards if required */ |
| 277 | while (target > bzf->bzf_bzstream.total_out_lo32) { |
| 278 | if (bzf_read(f, discard, min(sizeof(discard), target - bzf->bzf_bzstream.total_out_lo32), NULL) == -1) |
| 279 | return(-1); |
| 280 | } |
| 281 | /* This is where we are (be honest if we overshot) */ |
| 282 | return (bzf->bzf_bzstream.total_out_lo32); |
| 283 | } |
| 284 | |
| 285 | static int |
| 286 | bzf_stat(struct open_file *f, struct stat *sb) |
| 287 | { |
| 288 | struct bz_file *bzf = (struct bz_file *)f->f_fsdata; |
| 289 | int result; |
| 290 | |
| 291 | /* stat as normal, but indicate that size is unknown */ |
| 292 | if ((result = fstat(bzf->bzf_rawfd, sb)) == 0) |
| 293 | sb->st_size = -1; |
| 294 | return(result); |
| 295 | } |
| 296 | |
| 297 | void |
| 298 | bz_internal_error(int errorcode) |
| 299 | { |
| 300 | panic("bzipfs: critical error %d in bzip2 library occurred\n", errorcode); |
| 301 | } |