| 1 | #include <stdio.h> |
| 2 | #include <stdlib.h> |
| 3 | #include <string.h> |
| 4 | #include <limits.h> |
| 5 | #include <errno.h> |
| 6 | #include <assert.h> |
| 7 | #include <fcntl.h> |
| 8 | #include <unistd.h> |
| 9 | |
| 10 | #include "strntol.h" |
| 11 | #include "pattern.h" |
| 12 | #include "../minmax.h" |
| 13 | #include "../oslib/strcasestr.h" |
| 14 | #include "../oslib/strndup.h" |
| 15 | |
| 16 | /** |
| 17 | * parse_file() - parses binary file to fill buffer |
| 18 | * @beg - string input, extract filename from this |
| 19 | * @out - output buffer where parsed number should be put |
| 20 | * @out_len - length of the output buffer |
| 21 | * @filled - pointer where number of bytes successfully |
| 22 | * parsed will be put |
| 23 | * |
| 24 | * Returns the end pointer where parsing has been stopped. |
| 25 | * In case of parsing error or lack of bytes in output buffer |
| 26 | * NULL will be returned. |
| 27 | */ |
| 28 | static const char *parse_file(const char *beg, char *out, |
| 29 | unsigned int out_len, |
| 30 | unsigned int *filled) |
| 31 | { |
| 32 | const char *end; |
| 33 | char *file; |
| 34 | int fd; |
| 35 | ssize_t count; |
| 36 | |
| 37 | if (!out_len) |
| 38 | goto err_out; |
| 39 | |
| 40 | assert(*beg == '\''); |
| 41 | beg++; |
| 42 | end = strchr(beg, '\''); |
| 43 | if (!end) |
| 44 | goto err_out; |
| 45 | |
| 46 | file = strndup(beg, end - beg); |
| 47 | if (file == NULL) |
| 48 | goto err_out; |
| 49 | |
| 50 | fd = open(file, O_RDONLY); |
| 51 | if (fd < 0) |
| 52 | goto err_free_out; |
| 53 | |
| 54 | count = read(fd, out, out_len); |
| 55 | if (count == -1) |
| 56 | goto err_free_close_out; |
| 57 | |
| 58 | *filled = count; |
| 59 | close(fd); |
| 60 | free(file); |
| 61 | |
| 62 | /* Catch up quote */ |
| 63 | return end + 1; |
| 64 | |
| 65 | err_free_close_out: |
| 66 | close(fd); |
| 67 | err_free_out: |
| 68 | free(file); |
| 69 | err_out: |
| 70 | return NULL; |
| 71 | |
| 72 | } |
| 73 | |
| 74 | /** |
| 75 | * parse_string() - parses string in double quotes, like "abc" |
| 76 | * @beg - string input |
| 77 | * @out - output buffer where parsed number should be put |
| 78 | * @out_len - length of the output buffer |
| 79 | * @filled - pointer where number of bytes successfully |
| 80 | * parsed will be put |
| 81 | * |
| 82 | * Returns the end pointer where parsing has been stopped. |
| 83 | * In case of parsing error or lack of bytes in output buffer |
| 84 | * NULL will be returned. |
| 85 | */ |
| 86 | static const char *parse_string(const char *beg, char *out, |
| 87 | unsigned int out_len, |
| 88 | unsigned int *filled) |
| 89 | { |
| 90 | const char *end; |
| 91 | |
| 92 | if (!out_len) |
| 93 | return NULL; |
| 94 | |
| 95 | assert(*beg == '"'); |
| 96 | beg++; |
| 97 | end = strchr(beg, '"'); |
| 98 | if (!end) |
| 99 | return NULL; |
| 100 | if (end - beg > out_len) |
| 101 | return NULL; |
| 102 | |
| 103 | memcpy(out, beg, end - beg); |
| 104 | *filled = end - beg; |
| 105 | |
| 106 | /* Catch up quote */ |
| 107 | return end + 1; |
| 108 | } |
| 109 | |
| 110 | /** |
| 111 | * parse_number() - parses numbers |
| 112 | * @beg - string input |
| 113 | * @out - output buffer where parsed number should be put |
| 114 | * @out_len - length of the output buffer |
| 115 | * @filled - pointer where number of bytes successfully |
| 116 | * parsed will be put |
| 117 | * |
| 118 | * Supports decimals in the range [INT_MIN, INT_MAX] and |
| 119 | * hexidecimals of any size, which should be started with |
| 120 | * prefix 0x or 0X. |
| 121 | * |
| 122 | * Returns the end pointer where parsing has been stopped. |
| 123 | * In case of parsing error or lack of bytes in output buffer |
| 124 | * NULL will be returned. |
| 125 | */ |
| 126 | static const char *parse_number(const char *beg, char *out, |
| 127 | unsigned int out_len, |
| 128 | unsigned int *filled) |
| 129 | { |
| 130 | const char *end; |
| 131 | unsigned int val; |
| 132 | long lval; |
| 133 | int num, i; |
| 134 | |
| 135 | if (!out_len) |
| 136 | return NULL; |
| 137 | |
| 138 | num = 0; |
| 139 | sscanf(beg, "0%*[xX]%*[0-9a-fA-F]%n", &num); |
| 140 | if (num == 0) { |
| 141 | /* Here we are trying to parse decimal */ |
| 142 | |
| 143 | char *_end; |
| 144 | |
| 145 | /* Looking ahead */ |
| 146 | _end = strcasestr(beg, "0x"); |
| 147 | if (_end) |
| 148 | num = _end - beg; |
| 149 | if (num) |
| 150 | lval = strntol(beg, num, &_end, 10); |
| 151 | else |
| 152 | lval = strtol(beg, &_end, 10); |
| 153 | if (beg == _end || lval > INT_MAX || lval < INT_MIN) |
| 154 | return NULL; |
| 155 | end = _end; |
| 156 | i = 0; |
| 157 | if (!lval) { |
| 158 | num = 0; |
| 159 | out[i] = 0x00; |
| 160 | i = 1; |
| 161 | } else { |
| 162 | val = (unsigned int)lval; |
| 163 | for (; val && out_len; out_len--, i++, val >>= 8) |
| 164 | out[i] = val & 0xff; |
| 165 | if (val) |
| 166 | return NULL; |
| 167 | } |
| 168 | } else { |
| 169 | assert(num > 2); |
| 170 | |
| 171 | /* Catch up 0x prefix */ |
| 172 | num -= 2; |
| 173 | beg += 2; |
| 174 | |
| 175 | /* Look back, handle this combined string: 0xff0x14 */ |
| 176 | if (beg[num] && !strncasecmp(&beg[num - 1], "0x", 2)) |
| 177 | num--; |
| 178 | |
| 179 | end = beg + num; |
| 180 | |
| 181 | for (i = 0; num && out_len; |
| 182 | out_len--, i++, num -= 2, beg += 2) { |
| 183 | const char *fmt; |
| 184 | |
| 185 | fmt = (num & 1 ? "%1hhx" : "%2hhx"); |
| 186 | sscanf(beg, fmt, &out[i]); |
| 187 | if (num & 1) { |
| 188 | num++; |
| 189 | beg--; |
| 190 | } |
| 191 | } |
| 192 | if (num) |
| 193 | return NULL; |
| 194 | } |
| 195 | |
| 196 | *filled = i; |
| 197 | return end; |
| 198 | |
| 199 | } |
| 200 | |
| 201 | /** |
| 202 | * parse_format() - parses formats, like %o, etc |
| 203 | * @in - string input |
| 204 | * @out - output buffer where space for format should be reserved |
| 205 | * @parsed - number of bytes which were already parsed so far |
| 206 | * @out_len - length of the output buffer |
| 207 | * @fmt_desc - format descritor array, what we expect to find |
| 208 | * @fmt_desc_sz - size of the format descritor array |
| 209 | * @fmt - format array, the output |
| 210 | * @fmt_sz - size of format array |
| 211 | * |
| 212 | * This function tries to find formats, e.g.: |
| 213 | * %o - offset of the block |
| 214 | * |
| 215 | * In case of successfull parsing it fills the format param |
| 216 | * with proper offset and the size of the expected value, which |
| 217 | * should be pasted into buffer using the format 'func' callback. |
| 218 | * |
| 219 | * Returns the end pointer where parsing has been stopped. |
| 220 | * In case of parsing error or lack of bytes in output buffer |
| 221 | * NULL will be returned. |
| 222 | */ |
| 223 | static const char *parse_format(const char *in, char *out, unsigned int parsed, |
| 224 | unsigned int out_len, unsigned int *filled, |
| 225 | const struct pattern_fmt_desc *fmt_desc, |
| 226 | unsigned int fmt_desc_sz, |
| 227 | struct pattern_fmt *fmt, unsigned int fmt_sz) |
| 228 | { |
| 229 | int i; |
| 230 | struct pattern_fmt *f = NULL; |
| 231 | unsigned int len = 0; |
| 232 | |
| 233 | if (!out_len || !fmt_desc || !fmt_desc_sz || !fmt || !fmt_sz) |
| 234 | return NULL; |
| 235 | |
| 236 | assert(*in == '%'); |
| 237 | |
| 238 | for (i = 0; i < fmt_desc_sz; i++) { |
| 239 | const struct pattern_fmt_desc *desc; |
| 240 | |
| 241 | desc = &fmt_desc[i]; |
| 242 | len = strlen(desc->fmt); |
| 243 | if (0 == strncmp(in, desc->fmt, len)) { |
| 244 | fmt->desc = desc; |
| 245 | fmt->off = parsed; |
| 246 | f = fmt; |
| 247 | break; |
| 248 | } |
| 249 | } |
| 250 | |
| 251 | if (!f) |
| 252 | return NULL; |
| 253 | if (f->desc->len > out_len) |
| 254 | return NULL; |
| 255 | |
| 256 | memset(out, '\0', f->desc->len); |
| 257 | *filled = f->desc->len; |
| 258 | |
| 259 | return in + len; |
| 260 | } |
| 261 | |
| 262 | /** |
| 263 | * parse_and_fill_pattern() - Parses combined input, which consists of strings, |
| 264 | * numbers and pattern formats. |
| 265 | * @in - string input |
| 266 | * @in_len - size of the input string |
| 267 | * @out - output buffer where parsed result will be put |
| 268 | * @out_len - lengths of the output buffer |
| 269 | * @fmt_desc - array of pattern format descriptors [input] |
| 270 | * @fmt_desc_sz - size of the format descriptor array |
| 271 | * @fmt - array of pattern formats [output] |
| 272 | * @fmt_sz - pointer where the size of pattern formats array stored [input], |
| 273 | * after successfull parsing this pointer will contain the number |
| 274 | * of parsed formats if any [output]. |
| 275 | * |
| 276 | * strings: |
| 277 | * bytes sequence in double quotes, e.g. "123". |
| 278 | * NOTE: there is no way to escape quote, so "123\"abc" does not work. |
| 279 | * |
| 280 | * numbers: |
| 281 | * hexidecimal - sequence of hex bytes starting from 0x or 0X prefix, |
| 282 | * e.g. 0xff12ceff1100ff |
| 283 | * decimal - decimal number in range [INT_MIN, INT_MAX] |
| 284 | * |
| 285 | * formats: |
| 286 | * %o - offset of block, reserved 8 bytes. |
| 287 | * |
| 288 | * Explicit examples of combined string: |
| 289 | * #1 #2 #3 #4 |
| 290 | * in="abcd" in=-1024 in=66 in=0xFF0X1 |
| 291 | * out=61 62 63 64 out=00 fc ff ff out=42 out=ff 01 |
| 292 | * |
| 293 | * #5 #6 |
| 294 | * in=%o in="123"0xFFeeCC |
| 295 | * out=00 00 00 00 00 00 00 00 out=31 32 33 ff ec cc |
| 296 | * |
| 297 | * #7 |
| 298 | * in=-100xab"1"%o"2" |
| 299 | * out=f6 ff ff ff ab 31 00 00 00 00 00 00 00 00 32 |
| 300 | * |
| 301 | * #9 |
| 302 | * in=%o0xdeadbeef%o |
| 303 | * out=00 00 00 00 00 00 00 00 de ad be ef 00 00 00 00 00 00 00 00 |
| 304 | * |
| 305 | * #10 |
| 306 | * in=0xfefefefefefefefefefefefefefefefefefefefefe |
| 307 | * out=fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe |
| 308 | * |
| 309 | * Returns number of bytes filled or err < 0 in case of failure. |
| 310 | */ |
| 311 | int parse_and_fill_pattern(const char *in, unsigned int in_len, |
| 312 | char *out, unsigned int out_len, |
| 313 | const struct pattern_fmt_desc *fmt_desc, |
| 314 | unsigned int fmt_desc_sz, |
| 315 | struct pattern_fmt *fmt, |
| 316 | unsigned int *fmt_sz_out) |
| 317 | { |
| 318 | const char *beg, *end, *out_beg = out; |
| 319 | unsigned int total = 0, fmt_rem = 0; |
| 320 | |
| 321 | if (!in || !in_len || !out || !out_len) |
| 322 | return -EINVAL; |
| 323 | if (fmt_sz_out) |
| 324 | fmt_rem = *fmt_sz_out; |
| 325 | |
| 326 | beg = in; |
| 327 | do { |
| 328 | unsigned int filled; |
| 329 | int parsed_fmt; |
| 330 | |
| 331 | filled = 0; |
| 332 | parsed_fmt = 0; |
| 333 | |
| 334 | switch (*beg) { |
| 335 | case '\'': |
| 336 | end = parse_file(beg, out, out_len, &filled); |
| 337 | break; |
| 338 | case '"': |
| 339 | end = parse_string(beg, out, out_len, &filled); |
| 340 | break; |
| 341 | case '%': |
| 342 | end = parse_format(beg, out, out - out_beg, out_len, |
| 343 | &filled, fmt_desc, fmt_desc_sz, |
| 344 | fmt, fmt_rem); |
| 345 | parsed_fmt = 1; |
| 346 | break; |
| 347 | default: |
| 348 | end = parse_number(beg, out, out_len, &filled); |
| 349 | break; |
| 350 | } |
| 351 | |
| 352 | if (!end) |
| 353 | return -EINVAL; |
| 354 | |
| 355 | if (parsed_fmt) { |
| 356 | assert(fmt_rem); |
| 357 | fmt_rem--; |
| 358 | fmt++; |
| 359 | } |
| 360 | |
| 361 | assert(end - beg <= in_len); |
| 362 | in_len -= end - beg; |
| 363 | beg = end; |
| 364 | |
| 365 | assert(filled); |
| 366 | assert(filled <= out_len); |
| 367 | out_len -= filled; |
| 368 | out += filled; |
| 369 | total += filled; |
| 370 | |
| 371 | } while (in_len); |
| 372 | |
| 373 | if (fmt_sz_out) |
| 374 | *fmt_sz_out -= fmt_rem; |
| 375 | return total; |
| 376 | } |
| 377 | |
| 378 | /** |
| 379 | * dup_pattern() - Duplicates part of the pattern all over the buffer. |
| 380 | * |
| 381 | * Returns 0 in case of success or errno < 0 in case of failure. |
| 382 | */ |
| 383 | static int dup_pattern(char *out, unsigned int out_len, unsigned int pattern_len) |
| 384 | { |
| 385 | unsigned int left, len, off; |
| 386 | |
| 387 | if (out_len <= pattern_len) |
| 388 | /* Normal case */ |
| 389 | return 0; |
| 390 | |
| 391 | off = pattern_len; |
| 392 | left = (out_len - off); |
| 393 | len = min(left, off); |
| 394 | |
| 395 | /* Duplicate leftover */ |
| 396 | while (left) { |
| 397 | memcpy(out + off, out, len); |
| 398 | left -= len; |
| 399 | off <<= 1; |
| 400 | len = min(left, off); |
| 401 | } |
| 402 | |
| 403 | return 0; |
| 404 | } |
| 405 | |
| 406 | /** |
| 407 | * cpy_pattern() - Copies pattern to the buffer. |
| 408 | * |
| 409 | * Function copies pattern along the whole buffer. |
| 410 | * |
| 411 | * Returns 0 in case of success or errno < 0 in case of failure. |
| 412 | */ |
| 413 | int cpy_pattern(const char *pattern, unsigned int pattern_len, |
| 414 | char *out, unsigned int out_len) |
| 415 | { |
| 416 | unsigned int len; |
| 417 | |
| 418 | if (!pattern || !pattern_len || !out || !out_len) |
| 419 | return -EINVAL; |
| 420 | |
| 421 | /* Copy pattern */ |
| 422 | len = min(pattern_len, out_len); |
| 423 | memcpy(out, pattern, len); |
| 424 | |
| 425 | /* Spread filled chunk all over the buffer */ |
| 426 | return dup_pattern(out, out_len, pattern_len); |
| 427 | } |
| 428 | |
| 429 | /** |
| 430 | * cmp_pattern() - Compares pattern and buffer. |
| 431 | * |
| 432 | * For the sake of performance this function avoids any loops. |
| 433 | * Firstly it tries to compare the buffer itself, checking that |
| 434 | * buffer consists of repeating patterns along the buffer size. |
| 435 | * |
| 436 | * If the difference is not found then the function tries to compare |
| 437 | * buffer and pattern. |
| 438 | * |
| 439 | * Returns 0 in case of success or errno < 0 in case of failure. |
| 440 | */ |
| 441 | int cmp_pattern(const char *pattern, unsigned int pattern_size, |
| 442 | unsigned int off, const char *buf, unsigned int len) |
| 443 | { |
| 444 | int rc; |
| 445 | unsigned int size; |
| 446 | |
| 447 | /* Find the difference in buffer */ |
| 448 | if (len > pattern_size) { |
| 449 | rc = memcmp(buf, buf + pattern_size, len - pattern_size); |
| 450 | if (rc) |
| 451 | return -EILSEQ; |
| 452 | } |
| 453 | /* Compare second part of the pattern with buffer */ |
| 454 | if (off) { |
| 455 | size = min(len, pattern_size - off); |
| 456 | rc = memcmp(buf, pattern + off, size); |
| 457 | if (rc) |
| 458 | return -EILSEQ; |
| 459 | buf += size; |
| 460 | len -= size; |
| 461 | } |
| 462 | /* Compare first part of the pattern or the whole pattern |
| 463 | * with buffer */ |
| 464 | if (len) { |
| 465 | size = min(len, (off ? off : pattern_size)); |
| 466 | rc = memcmp(buf, pattern, size); |
| 467 | if (rc) |
| 468 | return -EILSEQ; |
| 469 | } |
| 470 | |
| 471 | return 0; |
| 472 | } |
| 473 | |
| 474 | /** |
| 475 | * paste_format_inplace() - Pastes parsed formats to the pattern. |
| 476 | * |
| 477 | * This function pastes formats to the pattern. If @fmt_sz is 0 |
| 478 | * function does nothing and pattern buffer is left untouched. |
| 479 | * |
| 480 | * Returns 0 in case of success or errno < 0 in case of failure. |
| 481 | */ |
| 482 | int paste_format_inplace(char *pattern, unsigned int pattern_len, |
| 483 | struct pattern_fmt *fmt, unsigned int fmt_sz, |
| 484 | void *priv) |
| 485 | { |
| 486 | int i, rc; |
| 487 | unsigned int len; |
| 488 | |
| 489 | if (!pattern || !pattern_len || !fmt) |
| 490 | return -EINVAL; |
| 491 | |
| 492 | /* Paste formats for first pattern chunk */ |
| 493 | for (i = 0; i < fmt_sz; i++) { |
| 494 | struct pattern_fmt *f; |
| 495 | |
| 496 | f = &fmt[i]; |
| 497 | if (pattern_len <= f->off) |
| 498 | break; |
| 499 | len = min(pattern_len - f->off, f->desc->len); |
| 500 | rc = f->desc->paste(pattern + f->off, len, priv); |
| 501 | if (rc) |
| 502 | return rc; |
| 503 | } |
| 504 | |
| 505 | return 0; |
| 506 | } |
| 507 | |
| 508 | /** |
| 509 | * paste_format() - Pastes parsed formats to the buffer. |
| 510 | * |
| 511 | * This function copies pattern to the buffer, pastes format |
| 512 | * into it and then duplicates pattern all over the buffer size. |
| 513 | * |
| 514 | * Returns 0 in case of success or errno < 0 in case of failure. |
| 515 | */ |
| 516 | int paste_format(const char *pattern, unsigned int pattern_len, |
| 517 | struct pattern_fmt *fmt, unsigned int fmt_sz, |
| 518 | char *out, unsigned int out_len, void *priv) |
| 519 | { |
| 520 | int rc; |
| 521 | unsigned int len; |
| 522 | |
| 523 | if (!pattern || !pattern_len || !out || !out_len) |
| 524 | return -EINVAL; |
| 525 | |
| 526 | /* Copy pattern */ |
| 527 | len = min(pattern_len, out_len); |
| 528 | memcpy(out, pattern, len); |
| 529 | |
| 530 | rc = paste_format_inplace(out, len, fmt, fmt_sz, priv); |
| 531 | if (rc) |
| 532 | return rc; |
| 533 | |
| 534 | /* Spread filled chunk all over the buffer */ |
| 535 | return dup_pattern(out, out_len, pattern_len); |
| 536 | } |