1/* $FreeBSD$ */ 2/* $NetBSD: unpack.c,v 1.3 2017/08/04 07:27:08 mrg Exp $ */ 3 4/*- 5 * SPDX-License-Identifier: BSD-2-Clause-FreeBSD 6 * 7 * Copyright (c) 2009 Xin LI <delphij@FreeBSD.org> 8 * 9 * Redistribution and use in source and binary forms, with or without 10 * modification, are permitted provided that the following conditions 11 * are met: 12 * 1. Redistributions of source code must retain the above copyright 13 * notice, this list of conditions and the following disclaimer. 14 * 2. Redistributions in binary form must reproduce the above copyright 15 * notice, this list of conditions and the following disclaimer in the 16 * documentation and/or other materials provided with the distribution. 17 * 18 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 19 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 20 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 21 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 22 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 23 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 24 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 25 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 26 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 27 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 28 * SUCH DAMAGE. 29 * 30 * $FreeBSD$ 31 */ 32 33/* This file is #included by gzip.c */ 34 35/* 36 * pack(1) file format: 37 * 38 * The first 7 bytes is the header: 39 * 00, 01 - Signature (US, RS), we already validated it earlier. 40 * 02..05 - Uncompressed size 41 * 06 - Level for the huffman tree (<=24) 42 * 43 * pack(1) will then store symbols (leaf) nodes count in each huffman 44 * tree levels, each level would consume 1 byte (See [1]). 45 * 46 * After the symbol count table, there is the symbol table, storing 47 * symbols represented by corresponding leaf node. EOB is not being 48 * explicitly transmitted (not necessary anyway) in the symbol table. 49 * 50 * Compressed data goes after the symbol table. 51 * 52 * NOTES 53 * 54 * [1] If we count EOB into the symbols, that would mean that we will 55 * have at most 256 symbols in the huffman tree. pack(1) rejects empty 56 * file and files that just repeats one character, which means that we 57 * will have at least 2 symbols. Therefore, pack(1) would reduce the 58 * last level symbol count by 2 which makes it a number in 59 * range [0..254], so all levels' symbol count would fit into 1 byte. 60 */ 61 62#define PACK_HEADER_LENGTH 7 63#define HTREE_MAXLEVEL 24 64 65/* 66 * unpack descriptor 67 * 68 * Represent the huffman tree in a similar way that pack(1) would 69 * store in a packed file. We store all symbols in a linear table, 70 * and store pointers to each level's first symbol. In addition to 71 * that, maintain two counts for each level: inner nodes count and 72 * leaf nodes count. 73 */ 74typedef struct { 75 int symbol_size; /* Size of the symbol table */ 76 int treelevels; /* Levels for the huffman tree */ 77 78 int *symbolsin; /* Table of leaf symbols count in each 79 * level */ 80 int *inodesin; /* Table of internal nodes count in 81 * each level */ 82 83 char *symbol; /* The symbol table */ 84 char *symbol_eob; /* Pointer to the EOB symbol */ 85 char **tree; /* Decoding huffman tree (pointers to 86 * first symbol of each tree level */ 87 88 off_t uncompressed_size; /* Uncompressed size */ 89 FILE *fpIn; /* Input stream */ 90 FILE *fpOut; /* Output stream */ 91} unpack_descriptor_t; 92 93/* 94 * Release resource allocated to an unpack descriptor. 95 * 96 * Caller is responsible to make sure that all of these pointers are 97 * initialized (in our case, they all point to valid memory block). 98 * We don't zero out pointers here because nobody else would ever 99 * reference the memory block without scrubbing them. 100 */ 101static void 102unpack_descriptor_fini(unpack_descriptor_t *unpackd) 103{ 104 105 free(unpackd->symbolsin); 106 free(unpackd->inodesin); 107 free(unpackd->symbol); 108 free(unpackd->tree); 109 110 fclose(unpackd->fpIn); 111 fclose(unpackd->fpOut); 112} 113 114/* 115 * Recursively fill the internal node count table 116 */ 117static void 118unpackd_fill_inodesin(const unpack_descriptor_t *unpackd, int level) 119{ 120 121 /* 122 * The internal nodes would be 1/2 of total internal nodes and 123 * leaf nodes in the next level. For the last level there 124 * would be no internal node by definition. 125 */ 126 if (level < unpackd->treelevels) { 127 unpackd_fill_inodesin(unpackd, level + 1); 128 unpackd->inodesin[level] = (unpackd->inodesin[level + 1] + 129 unpackd->symbolsin[level + 1]) / 2; 130 } else 131 unpackd->inodesin[level] = 0; 132} 133 134/* 135 * Update counter for accepted bytes 136 */ 137static void 138accepted_bytes(off_t *bytes_in, off_t newbytes) 139{ 140 141 if (bytes_in != NULL) 142 (*bytes_in) += newbytes; 143} 144 145/* 146 * Read file header and construct the tree. Also, prepare the buffered I/O 147 * for decode routine. 148 * 149 * Return value is uncompressed size. 150 */ 151static void 152unpack_parse_header(int in, int out, char *pre, size_t prelen, off_t *bytes_in, 153 unpack_descriptor_t *unpackd) 154{ 155 unsigned char hdr[PACK_HEADER_LENGTH]; /* buffer for header */ 156 ssize_t bytesread; /* Bytes read from the file */ 157 int i, j, thisbyte; 158 159 /* Prepend the header buffer if we already read some data */ 160 if (prelen != 0) 161 memcpy(hdr, pre, prelen); 162 163 /* Read in and fill the rest bytes of header */ 164 bytesread = read(in, hdr + prelen, PACK_HEADER_LENGTH - prelen); 165 if (bytesread < 0) 166 maybe_err("Error reading pack header"); 167 infile_newdata(bytesread); 168 169 accepted_bytes(bytes_in, PACK_HEADER_LENGTH); 170 171 /* Obtain uncompressed length (bytes 2,3,4,5) */ 172 unpackd->uncompressed_size = 0; 173 for (i = 2; i <= 5; i++) { 174 unpackd->uncompressed_size <<= 8; 175 unpackd->uncompressed_size |= hdr[i]; 176 } 177 178 /* Get the levels of the tree */ 179 unpackd->treelevels = hdr[6]; 180 if (unpackd->treelevels > HTREE_MAXLEVEL || unpackd->treelevels < 1) 181 maybe_errx("Huffman tree has insane levels"); 182 183 /* Let libc take care for buffering from now on */ 184 if ((unpackd->fpIn = fdopen(in, "r")) == NULL) 185 maybe_err("Can not fdopen() input stream"); 186 if ((unpackd->fpOut = fdopen(out, "w")) == NULL) 187 maybe_err("Can not fdopen() output stream"); 188 189 /* Allocate for the tables of bounds and the tree itself */ 190 unpackd->inodesin = 191 calloc(unpackd->treelevels, sizeof(*(unpackd->inodesin))); 192 unpackd->symbolsin = 193 calloc(unpackd->treelevels, sizeof(*(unpackd->symbolsin))); 194 unpackd->tree = 195 calloc(unpackd->treelevels, (sizeof(*(unpackd->tree)))); 196 if (unpackd->inodesin == NULL || unpackd->symbolsin == NULL || 197 unpackd->tree == NULL) 198 maybe_err("calloc"); 199 200 /* We count from 0 so adjust to match array upper bound */ 201 unpackd->treelevels--; 202 203 /* Read the levels symbol count table and calculate total */ 204 unpackd->symbol_size = 1; /* EOB */ 205 for (i = 0; i <= unpackd->treelevels; i++) { 206 if ((thisbyte = fgetc(unpackd->fpIn)) == EOF) 207 maybe_err("File appears to be truncated"); 208 unpackd->symbolsin[i] = (unsigned char)thisbyte; 209 unpackd->symbol_size += unpackd->symbolsin[i]; 210 } 211 accepted_bytes(bytes_in, unpackd->treelevels); 212 if (unpackd->symbol_size > 256) 213 maybe_errx("Bad symbol table"); 214 infile_newdata(unpackd->treelevels); 215 216 /* Allocate for the symbol table, point symbol_eob at the beginning */ 217 unpackd->symbol_eob = unpackd->symbol = calloc(1, unpackd->symbol_size); 218 if (unpackd->symbol == NULL) 219 maybe_err("calloc"); 220 221 /* 222 * Read in the symbol table, which contain [2, 256] symbols. 223 * In order to fit the count in one byte, pack(1) would offset 224 * it by reducing 2 from the actual number from the last level. 225 * 226 * We adjust the last level's symbol count by 1 here, because 227 * the EOB symbol is not being transmitted explicitly. Another 228 * adjustment would be done later afterward. 229 */ 230 unpackd->symbolsin[unpackd->treelevels]++; 231 for (i = 0; i <= unpackd->treelevels; i++) { 232 unpackd->tree[i] = unpackd->symbol_eob; 233 for (j = 0; j < unpackd->symbolsin[i]; j++) { 234 if ((thisbyte = fgetc(unpackd->fpIn)) == EOF) 235 maybe_errx("Symbol table truncated"); 236 *unpackd->symbol_eob++ = (char)thisbyte; 237 } 238 infile_newdata(unpackd->symbolsin[i]); 239 accepted_bytes(bytes_in, unpackd->symbolsin[i]); 240 } 241 242 /* Now, take account for the EOB symbol as well */ 243 unpackd->symbolsin[unpackd->treelevels]++; 244 245 /* 246 * The symbolsin table has been constructed now. 247 * Calculate the internal nodes count table based on it. 248 */ 249 unpackd_fill_inodesin(unpackd, 0); 250} 251 252/* 253 * Decode huffman stream, based on the huffman tree. 254 */ 255static void 256unpack_decode(const unpack_descriptor_t *unpackd, off_t *bytes_in) 257{ 258 int thislevel, thiscode, thisbyte, inlevelindex; 259 int i; 260 off_t bytes_out = 0; 261 const char *thissymbol; /* The symbol pointer decoded from stream */ 262 263 /* 264 * Decode huffman. Fetch every bytes from the file, get it 265 * into 'thiscode' bit-by-bit, then output the symbol we got 266 * when one has been found. 267 * 268 * Assumption: sizeof(int) > ((max tree levels + 1) / 8). 269 * bad things could happen if not. 270 */ 271 thislevel = 0; 272 thiscode = thisbyte = 0; 273 274 while ((thisbyte = fgetc(unpackd->fpIn)) != EOF) { 275 accepted_bytes(bytes_in, 1); 276 infile_newdata(1); 277 check_siginfo(); 278 279 /* 280 * Split one bit from thisbyte, from highest to lowest, 281 * feed the bit into thiscode, until we got a symbol from 282 * the tree. 283 */ 284 for (i = 7; i >= 0; i--) { 285 thiscode = (thiscode << 1) | ((thisbyte >> i) & 1); 286 287 /* Did we got a symbol? (referencing leaf node) */ 288 if (thiscode >= unpackd->inodesin[thislevel]) { 289 inlevelindex = 290 thiscode - unpackd->inodesin[thislevel]; 291 if (inlevelindex > unpackd->symbolsin[thislevel]) 292 maybe_errx("File corrupt"); 293 294 thissymbol = 295 &(unpackd->tree[thislevel][inlevelindex]); 296 if ((thissymbol == unpackd->symbol_eob) && 297 (bytes_out == unpackd->uncompressed_size)) 298 goto finished; 299 300 fputc((*thissymbol), unpackd->fpOut); 301 bytes_out++; 302 303 /* Prepare for next input */ 304 thislevel = 0; thiscode = 0; 305 } else { 306 thislevel++; 307 if (thislevel > unpackd->treelevels) 308 maybe_errx("File corrupt"); 309 } 310 } 311 } 312 313finished: 314 if (bytes_out != unpackd->uncompressed_size) 315 maybe_errx("Premature EOF"); 316} 317 318/* Handler for pack(1)'ed file */ 319static off_t 320unpack(int in, int out, char *pre, size_t prelen, off_t *bytes_in) 321{ 322 unpack_descriptor_t unpackd; 323 324 in = dup(in); 325 if (in == -1) 326 maybe_err("dup"); 327 out = dup(out); 328 if (out == -1) 329 maybe_err("dup"); 330 331 unpack_parse_header(in, out, pre, prelen, bytes_in, &unpackd); 332 unpack_decode(&unpackd, bytes_in); 333 unpack_descriptor_fini(&unpackd); 334 335 /* If we reached here, the unpack was successful */ 336 return (unpackd.uncompressed_size); 337} 338