/libavformat/id3v2.c
C | 1278 lines | 1005 code | 186 blank | 87 comment | 233 complexity | 13cfc49dde12b6fcbaf840b806648c92 MD5 | raw file
1/* 2 * Copyright (c) 2003 Fabrice Bellard 3 * 4 * This file is part of FFmpeg. 5 * 6 * FFmpeg is free software; you can redistribute it and/or 7 * modify it under the terms of the GNU Lesser General Public 8 * License as published by the Free Software Foundation; either 9 * version 2.1 of the License, or (at your option) any later version. 10 * 11 * FFmpeg is distributed in the hope that it will be useful, 12 * but WITHOUT ANY WARRANTY; without even the implied warranty of 13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU 14 * Lesser General Public License for more details. 15 * 16 * You should have received a copy of the GNU Lesser General Public 17 * License along with FFmpeg; if not, write to the Free Software 18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA 19 */ 20 21/** 22 * @file 23 * ID3v2 header parser 24 * 25 * Specifications available at: 26 * http://id3.org/Developer_Information 27 */ 28 29#include "config.h" 30 31#if CONFIG_ZLIB 32#include <zlib.h> 33#endif 34 35#include "libavutil/avstring.h" 36#include "libavutil/bprint.h" 37#include "libavutil/dict.h" 38#include "libavutil/intreadwrite.h" 39#include "libavcodec/png.h" 40#include "avio_internal.h" 41#include "internal.h" 42#include "id3v1.h" 43#include "id3v2.h" 44 45const AVMetadataConv ff_id3v2_34_metadata_conv[] = { 46 { "TALB", "album" }, 47 { "TCOM", "composer" }, 48 { "TCON", "genre" }, 49 { "TCOP", "copyright" }, 50 { "TENC", "encoded_by" }, 51 { "TIT2", "title" }, 52 { "TLAN", "language" }, 53 { "TPE1", "artist" }, 54 { "TPE2", "album_artist" }, 55 { "TPE3", "performer" }, 56 { "TPOS", "disc" }, 57 { "TPUB", "publisher" }, 58 { "TRCK", "track" }, 59 { "TSSE", "encoder" }, 60 { "USLT", "lyrics" }, 61 { 0 } 62}; 63 64const AVMetadataConv ff_id3v2_4_metadata_conv[] = { 65 { "TCMP", "compilation" }, 66 { "TDRC", "date" }, 67 { "TDRL", "date" }, 68 { "TDEN", "creation_time" }, 69 { "TSOA", "album-sort" }, 70 { "TSOP", "artist-sort" }, 71 { "TSOT", "title-sort" }, 72 { 0 } 73}; 74 75static const AVMetadataConv id3v2_2_metadata_conv[] = { 76 { "TAL", "album" }, 77 { "TCO", "genre" }, 78 { "TCP", "compilation" }, 79 { "TT2", "title" }, 80 { "TEN", "encoded_by" }, 81 { "TP1", "artist" }, 82 { "TP2", "album_artist" }, 83 { "TP3", "performer" }, 84 { "TRK", "track" }, 85 { 0 } 86}; 87 88const char ff_id3v2_tags[][4] = { 89 "TALB", "TBPM", "TCOM", "TCON", "TCOP", "TDLY", "TENC", "TEXT", 90 "TFLT", "TIT1", "TIT2", "TIT3", "TKEY", "TLAN", "TLEN", "TMED", 91 "TOAL", "TOFN", "TOLY", "TOPE", "TOWN", "TPE1", "TPE2", "TPE3", 92 "TPE4", "TPOS", "TPUB", "TRCK", "TRSN", "TRSO", "TSRC", "TSSE", 93 { 0 }, 94}; 95 96const char ff_id3v2_4_tags[][4] = { 97 "TDEN", "TDOR", "TDRC", "TDRL", "TDTG", "TIPL", "TMCL", "TMOO", 98 "TPRO", "TSOA", "TSOP", "TSOT", "TSST", 99 { 0 }, 100}; 101 102const char ff_id3v2_3_tags[][4] = { 103 "TDAT", "TIME", "TORY", "TRDA", "TSIZ", "TYER", 104 { 0 }, 105}; 106 107const char * const ff_id3v2_picture_types[21] = { 108 "Other", 109 "32x32 pixels 'file icon'", 110 "Other file icon", 111 "Cover (front)", 112 "Cover (back)", 113 "Leaflet page", 114 "Media (e.g. label side of CD)", 115 "Lead artist/lead performer/soloist", 116 "Artist/performer", 117 "Conductor", 118 "Band/Orchestra", 119 "Composer", 120 "Lyricist/text writer", 121 "Recording Location", 122 "During recording", 123 "During performance", 124 "Movie/video screen capture", 125 "A bright coloured fish", 126 "Illustration", 127 "Band/artist logotype", 128 "Publisher/Studio logotype", 129}; 130 131const CodecMime ff_id3v2_mime_tags[] = { 132 { "image/gif", AV_CODEC_ID_GIF }, 133 { "image/jpeg", AV_CODEC_ID_MJPEG }, 134 { "image/jpg", AV_CODEC_ID_MJPEG }, 135 { "image/png", AV_CODEC_ID_PNG }, 136 { "image/tiff", AV_CODEC_ID_TIFF }, 137 { "image/bmp", AV_CODEC_ID_BMP }, 138 { "JPG", AV_CODEC_ID_MJPEG }, /* ID3v2.2 */ 139 { "PNG", AV_CODEC_ID_PNG }, /* ID3v2.2 */ 140 { "", AV_CODEC_ID_NONE }, 141}; 142 143int ff_id3v2_match(const uint8_t *buf, const char *magic) 144{ 145 return buf[0] == magic[0] && 146 buf[1] == magic[1] && 147 buf[2] == magic[2] && 148 buf[3] != 0xff && 149 buf[4] != 0xff && 150 (buf[6] & 0x80) == 0 && 151 (buf[7] & 0x80) == 0 && 152 (buf[8] & 0x80) == 0 && 153 (buf[9] & 0x80) == 0; 154} 155 156int ff_id3v2_tag_len(const uint8_t *buf) 157{ 158 int len = ((buf[6] & 0x7f) << 21) + 159 ((buf[7] & 0x7f) << 14) + 160 ((buf[8] & 0x7f) << 7) + 161 (buf[9] & 0x7f) + 162 ID3v2_HEADER_SIZE; 163 if (buf[5] & 0x10) 164 len += ID3v2_HEADER_SIZE; 165 return len; 166} 167 168static unsigned int get_size(AVIOContext *s, int len) 169{ 170 int v = 0; 171 while (len--) 172 v = (v << 7) + (avio_r8(s) & 0x7F); 173 return v; 174} 175 176static unsigned int size_to_syncsafe(unsigned int size) 177{ 178 return (((size) & (0x7f << 0)) >> 0) + 179 (((size) & (0x7f << 8)) >> 1) + 180 (((size) & (0x7f << 16)) >> 2) + 181 (((size) & (0x7f << 24)) >> 3); 182} 183 184/* No real verification, only check that the tag consists of 185 * a combination of capital alpha-numerical characters */ 186static int is_tag(const char *buf, unsigned int len) 187{ 188 if (!len) 189 return 0; 190 191 while (len--) 192 if ((buf[len] < 'A' || 193 buf[len] > 'Z') && 194 (buf[len] < '0' || 195 buf[len] > '9')) 196 return 0; 197 198 return 1; 199} 200 201/** 202 * Return 1 if the tag of length len at the given offset is valid, 0 if not, -1 on error 203 */ 204static int check_tag(AVIOContext *s, int offset, unsigned int len) 205{ 206 char tag[4]; 207 208 if (len > 4 || 209 avio_seek(s, offset, SEEK_SET) < 0 || 210 avio_read(s, tag, len) < (int)len) 211 return -1; 212 else if (!AV_RB32(tag) || is_tag(tag, len)) 213 return 1; 214 215 return 0; 216} 217 218/** 219 * Free GEOB type extra metadata. 220 */ 221static void free_geobtag(void *obj) 222{ 223 ID3v2ExtraMetaGEOB *geob = obj; 224 av_freep(&geob->mime_type); 225 av_freep(&geob->file_name); 226 av_freep(&geob->description); 227 av_freep(&geob->data); 228 av_free(geob); 229} 230 231/** 232 * Decode characters to UTF-8 according to encoding type. The decoded buffer is 233 * always null terminated. Stop reading when either *maxread bytes are read from 234 * pb or U+0000 character is found. 235 * 236 * @param dst Pointer where the address of the buffer with the decoded bytes is 237 * stored. Buffer must be freed by caller. 238 * @param maxread Pointer to maximum number of characters to read from the 239 * AVIOContext. After execution the value is decremented by the number of bytes 240 * actually read. 241 * @returns 0 if no error occurred, dst is uninitialized on error 242 */ 243static int decode_str(AVFormatContext *s, AVIOContext *pb, int encoding, 244 uint8_t **dst, int *maxread) 245{ 246 int ret; 247 uint8_t tmp; 248 uint32_t ch = 1; 249 int left = *maxread; 250 unsigned int (*get)(AVIOContext*) = avio_rb16; 251 AVIOContext *dynbuf; 252 253 if ((ret = avio_open_dyn_buf(&dynbuf)) < 0) { 254 av_log(s, AV_LOG_ERROR, "Error opening memory stream\n"); 255 return ret; 256 } 257 258 switch (encoding) { 259 case ID3v2_ENCODING_ISO8859: 260 while (left && ch) { 261 ch = avio_r8(pb); 262 PUT_UTF8(ch, tmp, avio_w8(dynbuf, tmp);) 263 left--; 264 } 265 break; 266 267 case ID3v2_ENCODING_UTF16BOM: 268 if ((left -= 2) < 0) { 269 av_log(s, AV_LOG_ERROR, "Cannot read BOM value, input too short\n"); 270 ffio_free_dyn_buf(&dynbuf); 271 *dst = NULL; 272 return AVERROR_INVALIDDATA; 273 } 274 switch (avio_rb16(pb)) { 275 case 0xfffe: 276 get = avio_rl16; 277 case 0xfeff: 278 break; 279 default: 280 av_log(s, AV_LOG_ERROR, "Incorrect BOM value\n"); 281 ffio_free_dyn_buf(&dynbuf); 282 *dst = NULL; 283 *maxread = left; 284 return AVERROR_INVALIDDATA; 285 } 286 // fall-through 287 288 case ID3v2_ENCODING_UTF16BE: 289 while ((left > 1) && ch) { 290 GET_UTF16(ch, ((left -= 2) >= 0 ? get(pb) : 0), break;) 291 PUT_UTF8(ch, tmp, avio_w8(dynbuf, tmp);) 292 } 293 if (left < 0) 294 left += 2; /* did not read last char from pb */ 295 break; 296 297 case ID3v2_ENCODING_UTF8: 298 while (left && ch) { 299 ch = avio_r8(pb); 300 avio_w8(dynbuf, ch); 301 left--; 302 } 303 break; 304 default: 305 av_log(s, AV_LOG_WARNING, "Unknown encoding\n"); 306 } 307 308 if (ch) 309 avio_w8(dynbuf, 0); 310 311 avio_close_dyn_buf(dynbuf, dst); 312 *maxread = left; 313 314 return 0; 315} 316 317/** 318 * Parse a text tag. 319 */ 320static void read_ttag(AVFormatContext *s, AVIOContext *pb, int taglen, 321 AVDictionary **metadata, const char *key) 322{ 323 uint8_t *dst; 324 int encoding, dict_flags = AV_DICT_DONT_OVERWRITE | AV_DICT_DONT_STRDUP_VAL; 325 unsigned genre; 326 327 if (taglen < 1) 328 return; 329 330 encoding = avio_r8(pb); 331 taglen--; /* account for encoding type byte */ 332 333 if (decode_str(s, pb, encoding, &dst, &taglen) < 0) { 334 av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", key); 335 return; 336 } 337 338 if (!(strcmp(key, "TCON") && strcmp(key, "TCO")) && 339 (sscanf(dst, "(%d)", &genre) == 1 || sscanf(dst, "%d", &genre) == 1) && 340 genre <= ID3v1_GENRE_MAX) { 341 av_freep(&dst); 342 dst = av_strdup(ff_id3v1_genre_str[genre]); 343 } else if (!(strcmp(key, "TXXX") && strcmp(key, "TXX"))) { 344 /* dst now contains the key, need to get value */ 345 key = dst; 346 if (decode_str(s, pb, encoding, &dst, &taglen) < 0) { 347 av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", key); 348 av_freep(&key); 349 return; 350 } 351 dict_flags |= AV_DICT_DONT_STRDUP_KEY; 352 } else if (!*dst) 353 av_freep(&dst); 354 355 if (dst) 356 av_dict_set(metadata, key, dst, dict_flags); 357} 358 359static void read_uslt(AVFormatContext *s, AVIOContext *pb, int taglen, 360 AVDictionary **metadata) 361{ 362 uint8_t lang[4]; 363 uint8_t *descriptor = NULL; // 'Content descriptor' 364 uint8_t *text; 365 char *key; 366 int encoding; 367 int ok = 0; 368 369 if (taglen < 1) 370 goto error; 371 372 encoding = avio_r8(pb); 373 taglen--; 374 375 if (avio_read(pb, lang, 3) < 3) 376 goto error; 377 lang[3] = '\0'; 378 taglen -= 3; 379 380 if (decode_str(s, pb, encoding, &descriptor, &taglen) < 0) 381 goto error; 382 383 if (decode_str(s, pb, encoding, &text, &taglen) < 0) 384 goto error; 385 386 // FFmpeg does not support hierarchical metadata, so concatenate the keys. 387 key = av_asprintf("lyrics-%s%s%s", descriptor[0] ? (char *)descriptor : "", 388 descriptor[0] ? "-" : "", 389 lang); 390 if (!key) { 391 av_free(text); 392 goto error; 393 } 394 395 av_dict_set(metadata, key, text, 396 AV_DICT_DONT_STRDUP_KEY | AV_DICT_DONT_STRDUP_VAL); 397 398 ok = 1; 399error: 400 if (!ok) 401 av_log(s, AV_LOG_ERROR, "Error reading lyrics, skipped\n"); 402 av_free(descriptor); 403} 404 405/** 406 * Parse a comment tag. 407 */ 408static void read_comment(AVFormatContext *s, AVIOContext *pb, int taglen, 409 AVDictionary **metadata) 410{ 411 const char *key = "comment"; 412 uint8_t *dst; 413 int encoding, dict_flags = AV_DICT_DONT_OVERWRITE | AV_DICT_DONT_STRDUP_VAL; 414 av_unused int language; 415 416 if (taglen < 4) 417 return; 418 419 encoding = avio_r8(pb); 420 language = avio_rl24(pb); 421 taglen -= 4; 422 423 if (decode_str(s, pb, encoding, &dst, &taglen) < 0) { 424 av_log(s, AV_LOG_ERROR, "Error reading comment frame, skipped\n"); 425 return; 426 } 427 428 if (dst && !*dst) 429 av_freep(&dst); 430 431 if (dst) { 432 key = (const char *) dst; 433 dict_flags |= AV_DICT_DONT_STRDUP_KEY; 434 } 435 436 if (decode_str(s, pb, encoding, &dst, &taglen) < 0) { 437 av_log(s, AV_LOG_ERROR, "Error reading comment frame, skipped\n"); 438 if (dict_flags & AV_DICT_DONT_STRDUP_KEY) 439 av_freep((void*)&key); 440 return; 441 } 442 443 if (dst) 444 av_dict_set(metadata, key, (const char *) dst, dict_flags); 445} 446 447/** 448 * Parse GEOB tag into a ID3v2ExtraMetaGEOB struct. 449 */ 450static void read_geobtag(AVFormatContext *s, AVIOContext *pb, int taglen, 451 const char *tag, ID3v2ExtraMeta **extra_meta, 452 int isv34) 453{ 454 ID3v2ExtraMetaGEOB *geob_data = NULL; 455 ID3v2ExtraMeta *new_extra = NULL; 456 char encoding; 457 unsigned int len; 458 459 if (taglen < 1) 460 return; 461 462 geob_data = av_mallocz(sizeof(ID3v2ExtraMetaGEOB)); 463 if (!geob_data) { 464 av_log(s, AV_LOG_ERROR, "Failed to alloc %"SIZE_SPECIFIER" bytes\n", 465 sizeof(ID3v2ExtraMetaGEOB)); 466 return; 467 } 468 469 new_extra = av_mallocz(sizeof(ID3v2ExtraMeta)); 470 if (!new_extra) { 471 av_log(s, AV_LOG_ERROR, "Failed to alloc %"SIZE_SPECIFIER" bytes\n", 472 sizeof(ID3v2ExtraMeta)); 473 goto fail; 474 } 475 476 /* read encoding type byte */ 477 encoding = avio_r8(pb); 478 taglen--; 479 480 /* read MIME type (always ISO-8859) */ 481 if (decode_str(s, pb, ID3v2_ENCODING_ISO8859, &geob_data->mime_type, 482 &taglen) < 0 || 483 taglen <= 0) 484 goto fail; 485 486 /* read file name */ 487 if (decode_str(s, pb, encoding, &geob_data->file_name, &taglen) < 0 || 488 taglen <= 0) 489 goto fail; 490 491 /* read content description */ 492 if (decode_str(s, pb, encoding, &geob_data->description, &taglen) < 0 || 493 taglen < 0) 494 goto fail; 495 496 if (taglen) { 497 /* save encapsulated binary data */ 498 geob_data->data = av_malloc(taglen); 499 if (!geob_data->data) { 500 av_log(s, AV_LOG_ERROR, "Failed to alloc %d bytes\n", taglen); 501 goto fail; 502 } 503 if ((len = avio_read(pb, geob_data->data, taglen)) < taglen) 504 av_log(s, AV_LOG_WARNING, 505 "Error reading GEOB frame, data truncated.\n"); 506 geob_data->datasize = len; 507 } else { 508 geob_data->data = NULL; 509 geob_data->datasize = 0; 510 } 511 512 /* add data to the list */ 513 new_extra->tag = "GEOB"; 514 new_extra->data = geob_data; 515 new_extra->next = *extra_meta; 516 *extra_meta = new_extra; 517 518 return; 519 520fail: 521 av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", tag); 522 free_geobtag(geob_data); 523 av_free(new_extra); 524 return; 525} 526 527static int is_number(const char *str) 528{ 529 while (*str >= '0' && *str <= '9') 530 str++; 531 return !*str; 532} 533 534static AVDictionaryEntry *get_date_tag(AVDictionary *m, const char *tag) 535{ 536 AVDictionaryEntry *t; 537 if ((t = av_dict_get(m, tag, NULL, AV_DICT_MATCH_CASE)) && 538 strlen(t->value) == 4 && is_number(t->value)) 539 return t; 540 return NULL; 541} 542 543static void merge_date(AVDictionary **m) 544{ 545 AVDictionaryEntry *t; 546 char date[17] = { 0 }; // YYYY-MM-DD hh:mm 547 548 if (!(t = get_date_tag(*m, "TYER")) && 549 !(t = get_date_tag(*m, "TYE"))) 550 return; 551 av_strlcpy(date, t->value, 5); 552 av_dict_set(m, "TYER", NULL, 0); 553 av_dict_set(m, "TYE", NULL, 0); 554 555 if (!(t = get_date_tag(*m, "TDAT")) && 556 !(t = get_date_tag(*m, "TDA"))) 557 goto finish; 558 snprintf(date + 4, sizeof(date) - 4, "-%.2s-%.2s", t->value + 2, t->value); 559 av_dict_set(m, "TDAT", NULL, 0); 560 av_dict_set(m, "TDA", NULL, 0); 561 562 if (!(t = get_date_tag(*m, "TIME")) && 563 !(t = get_date_tag(*m, "TIM"))) 564 goto finish; 565 snprintf(date + 10, sizeof(date) - 10, 566 " %.2s:%.2s", t->value, t->value + 2); 567 av_dict_set(m, "TIME", NULL, 0); 568 av_dict_set(m, "TIM", NULL, 0); 569 570finish: 571 if (date[0]) 572 av_dict_set(m, "date", date, 0); 573} 574 575static void free_apic(void *obj) 576{ 577 ID3v2ExtraMetaAPIC *apic = obj; 578 av_buffer_unref(&apic->buf); 579 av_freep(&apic->description); 580 av_freep(&apic); 581} 582 583static void rstrip_spaces(char *buf) 584{ 585 size_t len = strlen(buf); 586 while (len > 0 && buf[len - 1] == ' ') 587 buf[--len] = 0; 588} 589 590static void read_apic(AVFormatContext *s, AVIOContext *pb, int taglen, 591 const char *tag, ID3v2ExtraMeta **extra_meta, 592 int isv34) 593{ 594 int enc, pic_type; 595 char mimetype[64] = {0}; 596 const CodecMime *mime = ff_id3v2_mime_tags; 597 enum AVCodecID id = AV_CODEC_ID_NONE; 598 ID3v2ExtraMetaAPIC *apic = NULL; 599 ID3v2ExtraMeta *new_extra = NULL; 600 int64_t end = avio_tell(pb) + taglen; 601 602 if (taglen <= 4 || (!isv34 && taglen <= 6)) 603 goto fail; 604 605 new_extra = av_mallocz(sizeof(*new_extra)); 606 apic = av_mallocz(sizeof(*apic)); 607 if (!new_extra || !apic) 608 goto fail; 609 610 enc = avio_r8(pb); 611 taglen--; 612 613 /* mimetype */ 614 if (isv34) { 615 taglen -= avio_get_str(pb, taglen, mimetype, sizeof(mimetype)); 616 } else { 617 if (avio_read(pb, mimetype, 3) < 0) 618 goto fail; 619 620 mimetype[3] = 0; 621 taglen -= 3; 622 } 623 624 while (mime->id != AV_CODEC_ID_NONE) { 625 if (!av_strncasecmp(mime->str, mimetype, sizeof(mimetype))) { 626 id = mime->id; 627 break; 628 } 629 mime++; 630 } 631 if (id == AV_CODEC_ID_NONE) { 632 av_log(s, AV_LOG_WARNING, 633 "Unknown attached picture mimetype: %s, skipping.\n", mimetype); 634 goto fail; 635 } 636 apic->id = id; 637 638 /* picture type */ 639 pic_type = avio_r8(pb); 640 taglen--; 641 if (pic_type < 0 || pic_type >= FF_ARRAY_ELEMS(ff_id3v2_picture_types)) { 642 av_log(s, AV_LOG_WARNING, "Unknown attached picture type %d.\n", 643 pic_type); 644 pic_type = 0; 645 } 646 apic->type = ff_id3v2_picture_types[pic_type]; 647 648 /* description and picture data */ 649 if (decode_str(s, pb, enc, &apic->description, &taglen) < 0) { 650 av_log(s, AV_LOG_ERROR, 651 "Error decoding attached picture description.\n"); 652 goto fail; 653 } 654 655 apic->buf = av_buffer_alloc(taglen + AV_INPUT_BUFFER_PADDING_SIZE); 656 if (!apic->buf || !taglen || avio_read(pb, apic->buf->data, taglen) != taglen) 657 goto fail; 658 memset(apic->buf->data + taglen, 0, AV_INPUT_BUFFER_PADDING_SIZE); 659 660 new_extra->tag = "APIC"; 661 new_extra->data = apic; 662 new_extra->next = *extra_meta; 663 *extra_meta = new_extra; 664 665 // The description must be unique, and some ID3v2 tag writers add spaces 666 // to write several APIC entries with the same description. 667 rstrip_spaces(apic->description); 668 669 return; 670 671fail: 672 if (apic) 673 free_apic(apic); 674 av_freep(&new_extra); 675 avio_seek(pb, end, SEEK_SET); 676} 677 678static void free_chapter(void *obj) 679{ 680 ID3v2ExtraMetaCHAP *chap = obj; 681 av_freep(&chap->element_id); 682 av_dict_free(&chap->meta); 683 av_freep(&chap); 684} 685 686static void read_chapter(AVFormatContext *s, AVIOContext *pb, int len, const char *ttag, ID3v2ExtraMeta **extra_meta, int isv34) 687{ 688 int taglen; 689 char tag[5]; 690 ID3v2ExtraMeta *new_extra = NULL; 691 ID3v2ExtraMetaCHAP *chap = NULL; 692 693 new_extra = av_mallocz(sizeof(*new_extra)); 694 chap = av_mallocz(sizeof(*chap)); 695 696 if (!new_extra || !chap) 697 goto fail; 698 699 if (decode_str(s, pb, 0, &chap->element_id, &len) < 0) 700 goto fail; 701 702 if (len < 16) 703 goto fail; 704 705 chap->start = avio_rb32(pb); 706 chap->end = avio_rb32(pb); 707 avio_skip(pb, 8); 708 709 len -= 16; 710 while (len > 10) { 711 if (avio_read(pb, tag, 4) < 4) 712 goto fail; 713 tag[4] = 0; 714 taglen = avio_rb32(pb); 715 avio_skip(pb, 2); 716 len -= 10; 717 if (taglen < 0 || taglen > len) 718 goto fail; 719 if (tag[0] == 'T') 720 read_ttag(s, pb, taglen, &chap->meta, tag); 721 else 722 avio_skip(pb, taglen); 723 len -= taglen; 724 } 725 726 ff_metadata_conv(&chap->meta, NULL, ff_id3v2_34_metadata_conv); 727 ff_metadata_conv(&chap->meta, NULL, ff_id3v2_4_metadata_conv); 728 729 new_extra->tag = "CHAP"; 730 new_extra->data = chap; 731 new_extra->next = *extra_meta; 732 *extra_meta = new_extra; 733 734 return; 735 736fail: 737 if (chap) 738 free_chapter(chap); 739 av_freep(&new_extra); 740} 741 742static void free_priv(void *obj) 743{ 744 ID3v2ExtraMetaPRIV *priv = obj; 745 av_freep(&priv->owner); 746 av_freep(&priv->data); 747 av_freep(&priv); 748} 749 750static void read_priv(AVFormatContext *s, AVIOContext *pb, int taglen, 751 const char *tag, ID3v2ExtraMeta **extra_meta, int isv34) 752{ 753 ID3v2ExtraMeta *meta; 754 ID3v2ExtraMetaPRIV *priv; 755 756 meta = av_mallocz(sizeof(*meta)); 757 priv = av_mallocz(sizeof(*priv)); 758 759 if (!meta || !priv) 760 goto fail; 761 762 if (decode_str(s, pb, ID3v2_ENCODING_ISO8859, &priv->owner, &taglen) < 0) 763 goto fail; 764 765 priv->data = av_malloc(taglen); 766 if (!priv->data) 767 goto fail; 768 769 priv->datasize = taglen; 770 771 if (avio_read(pb, priv->data, priv->datasize) != priv->datasize) 772 goto fail; 773 774 meta->tag = "PRIV"; 775 meta->data = priv; 776 meta->next = *extra_meta; 777 *extra_meta = meta; 778 779 return; 780 781fail: 782 if (priv) 783 free_priv(priv); 784 av_freep(&meta); 785} 786 787typedef struct ID3v2EMFunc { 788 const char *tag3; 789 const char *tag4; 790 void (*read)(AVFormatContext *s, AVIOContext *pb, int taglen, 791 const char *tag, ID3v2ExtraMeta **extra_meta, 792 int isv34); 793 void (*free)(void *obj); 794} ID3v2EMFunc; 795 796static const ID3v2EMFunc id3v2_extra_meta_funcs[] = { 797 { "GEO", "GEOB", read_geobtag, free_geobtag }, 798 { "PIC", "APIC", read_apic, free_apic }, 799 { "CHAP","CHAP", read_chapter, free_chapter }, 800 { "PRIV","PRIV", read_priv, free_priv }, 801 { NULL } 802}; 803 804/** 805 * Get the corresponding ID3v2EMFunc struct for a tag. 806 * @param isv34 Determines if v2.2 or v2.3/4 strings are used 807 * @return A pointer to the ID3v2EMFunc struct if found, NULL otherwise. 808 */ 809static const ID3v2EMFunc *get_extra_meta_func(const char *tag, int isv34) 810{ 811 int i = 0; 812 while (id3v2_extra_meta_funcs[i].tag3) { 813 if (tag && !memcmp(tag, 814 (isv34 ? id3v2_extra_meta_funcs[i].tag4 : 815 id3v2_extra_meta_funcs[i].tag3), 816 (isv34 ? 4 : 3))) 817 return &id3v2_extra_meta_funcs[i]; 818 i++; 819 } 820 return NULL; 821} 822 823static void id3v2_parse(AVIOContext *pb, AVDictionary **metadata, 824 AVFormatContext *s, int len, uint8_t version, 825 uint8_t flags, ID3v2ExtraMeta **extra_meta) 826{ 827 int isv34, unsync; 828 unsigned tlen; 829 char tag[5]; 830 int64_t next, end = avio_tell(pb) + len; 831 int taghdrlen; 832 const char *reason = NULL; 833 AVIOContext pb_local; 834 AVIOContext *pbx; 835 unsigned char *buffer = NULL; 836 int buffer_size = 0; 837 const ID3v2EMFunc *extra_func = NULL; 838 unsigned char *uncompressed_buffer = NULL; 839 av_unused int uncompressed_buffer_size = 0; 840 const char *comm_frame; 841 842 av_log(s, AV_LOG_DEBUG, "id3v2 ver:%d flags:%02X len:%d\n", version, flags, len); 843 844 switch (version) { 845 case 2: 846 if (flags & 0x40) { 847 reason = "compression"; 848 goto error; 849 } 850 isv34 = 0; 851 taghdrlen = 6; 852 comm_frame = "COM"; 853 break; 854 855 case 3: 856 case 4: 857 isv34 = 1; 858 taghdrlen = 10; 859 comm_frame = "COMM"; 860 break; 861 862 default: 863 reason = "version"; 864 goto error; 865 } 866 867 unsync = flags & 0x80; 868 869 if (isv34 && flags & 0x40) { /* Extended header present, just skip over it */ 870 int extlen = get_size(pb, 4); 871 if (version == 4) 872 /* In v2.4 the length includes the length field we just read. */ 873 extlen -= 4; 874 875 if (extlen < 0) { 876 reason = "invalid extended header length"; 877 goto error; 878 } 879 avio_skip(pb, extlen); 880 len -= extlen + 4; 881 if (len < 0) { 882 reason = "extended header too long."; 883 goto error; 884 } 885 } 886 887 while (len >= taghdrlen) { 888 unsigned int tflags = 0; 889 int tunsync = 0; 890 int tcomp = 0; 891 int tencr = 0; 892 unsigned long av_unused dlen; 893 894 if (isv34) { 895 if (avio_read(pb, tag, 4) < 4) 896 break; 897 tag[4] = 0; 898 if (version == 3) { 899 tlen = avio_rb32(pb); 900 } else { 901 /* some encoders incorrectly uses v3 sizes instead of syncsafe ones 902 * so check the next tag to see which one to use */ 903 tlen = avio_rb32(pb); 904 if (tlen > 0x7f) { 905 if (tlen < len) { 906 int64_t cur = avio_tell(pb); 907 908 if (ffio_ensure_seekback(pb, 2 /* tflags */ + tlen + 4 /* next tag */)) 909 break; 910 911 if (check_tag(pb, cur + 2 + size_to_syncsafe(tlen), 4) == 1) 912 tlen = size_to_syncsafe(tlen); 913 else if (check_tag(pb, cur + 2 + tlen, 4) != 1) 914 break; 915 avio_seek(pb, cur, SEEK_SET); 916 } else 917 tlen = size_to_syncsafe(tlen); 918 } 919 } 920 tflags = avio_rb16(pb); 921 tunsync = tflags & ID3v2_FLAG_UNSYNCH; 922 } else { 923 if (avio_read(pb, tag, 3) < 3) 924 break; 925 tag[3] = 0; 926 tlen = avio_rb24(pb); 927 } 928 if (tlen > (1<<28)) 929 break; 930 len -= taghdrlen + tlen; 931 932 if (len < 0) 933 break; 934 935 next = avio_tell(pb) + tlen; 936 937 if (!tlen) { 938 if (tag[0]) 939 av_log(s, AV_LOG_DEBUG, "Invalid empty frame %s, skipping.\n", 940 tag); 941 continue; 942 } 943 944 if (tflags & ID3v2_FLAG_DATALEN) { 945 if (tlen < 4) 946 break; 947 dlen = avio_rb32(pb); 948 tlen -= 4; 949 } else 950 dlen = tlen; 951 952 tcomp = tflags & ID3v2_FLAG_COMPRESSION; 953 tencr = tflags & ID3v2_FLAG_ENCRYPTION; 954 955 /* skip encrypted tags and, if no zlib, compressed tags */ 956 if (tencr || (!CONFIG_ZLIB && tcomp)) { 957 const char *type; 958 if (!tcomp) 959 type = "encrypted"; 960 else if (!tencr) 961 type = "compressed"; 962 else 963 type = "encrypted and compressed"; 964 965 av_log(s, AV_LOG_WARNING, "Skipping %s ID3v2 frame %s.\n", type, tag); 966 avio_skip(pb, tlen); 967 /* check for text tag or supported special meta tag */ 968 } else if (tag[0] == 'T' || 969 !memcmp(tag, "USLT", 4) || 970 !strcmp(tag, comm_frame) || 971 (extra_meta && 972 (extra_func = get_extra_meta_func(tag, isv34)))) { 973 pbx = pb; 974 975 if (unsync || tunsync || tcomp) { 976 av_fast_malloc(&buffer, &buffer_size, tlen); 977 if (!buffer) { 978 av_log(s, AV_LOG_ERROR, "Failed to alloc %d bytes\n", tlen); 979 goto seek; 980 } 981 } 982 if (unsync || tunsync) { 983 uint8_t *b = buffer; 984 uint8_t *t = buffer; 985 uint8_t *end = t + tlen; 986 987 if (avio_read(pb, buffer, tlen) != tlen) { 988 av_log(s, AV_LOG_ERROR, "Failed to read tag data\n"); 989 goto seek; 990 } 991 992 while (t != end) { 993 *b++ = *t++; 994 if (t != end && t[-1] == 0xff && !t[0]) 995 t++; 996 } 997 998 ffio_init_context(&pb_local, buffer, b - buffer, 0, NULL, NULL, NULL, 999 NULL); 1000 tlen = b - buffer; 1001 pbx = &pb_local; // read from sync buffer 1002 } 1003 1004#if CONFIG_ZLIB 1005 if (tcomp) { 1006 int err; 1007 1008 av_log(s, AV_LOG_DEBUG, "Compresssed frame %s tlen=%d dlen=%ld\n", tag, tlen, dlen); 1009 1010 av_fast_malloc(&uncompressed_buffer, &uncompressed_buffer_size, dlen); 1011 if (!uncompressed_buffer) { 1012 av_log(s, AV_LOG_ERROR, "Failed to alloc %ld bytes\n", dlen); 1013 goto seek; 1014 } 1015 1016 if (!(unsync || tunsync)) { 1017 err = avio_read(pb, buffer, tlen); 1018 if (err < 0) { 1019 av_log(s, AV_LOG_ERROR, "Failed to read compressed tag\n"); 1020 goto seek; 1021 } 1022 tlen = err; 1023 } 1024 1025 err = uncompress(uncompressed_buffer, &dlen, buffer, tlen); 1026 if (err != Z_OK) { 1027 av_log(s, AV_LOG_ERROR, "Failed to uncompress tag: %d\n", err); 1028 goto seek; 1029 } 1030 ffio_init_context(&pb_local, uncompressed_buffer, dlen, 0, NULL, NULL, NULL, NULL); 1031 tlen = dlen; 1032 pbx = &pb_local; // read from sync buffer 1033 } 1034#endif 1035 if (tag[0] == 'T') 1036 /* parse text tag */ 1037 read_ttag(s, pbx, tlen, metadata, tag); 1038 else if (!memcmp(tag, "USLT", 4)) 1039 read_uslt(s, pbx, tlen, metadata); 1040 else if (!strcmp(tag, comm_frame)) 1041 read_comment(s, pbx, tlen, metadata); 1042 else 1043 /* parse special meta tag */ 1044 extra_func->read(s, pbx, tlen, tag, extra_meta, isv34); 1045 } else if (!tag[0]) { 1046 if (tag[1]) 1047 av_log(s, AV_LOG_WARNING, "invalid frame id, assuming padding\n"); 1048 avio_skip(pb, tlen); 1049 break; 1050 } 1051 /* Skip to end of tag */ 1052seek: 1053 avio_seek(pb, next, SEEK_SET); 1054 } 1055 1056 /* Footer preset, always 10 bytes, skip over it */ 1057 if (version == 4 && flags & 0x10) 1058 end += 10; 1059 1060error: 1061 if (reason) 1062 av_log(s, AV_LOG_INFO, "ID3v2.%d tag skipped, cannot handle %s\n", 1063 version, reason); 1064 avio_seek(pb, end, SEEK_SET); 1065 av_free(buffer); 1066 av_free(uncompressed_buffer); 1067 return; 1068} 1069 1070static void id3v2_read_internal(AVIOContext *pb, AVDictionary **metadata, 1071 AVFormatContext *s, const char *magic, 1072 ID3v2ExtraMeta **extra_meta, int64_t max_search_size) 1073{ 1074 int len, ret; 1075 uint8_t buf[ID3v2_HEADER_SIZE]; 1076 int found_header; 1077 int64_t start, off; 1078 1079 if (max_search_size && max_search_size < ID3v2_HEADER_SIZE) 1080 return; 1081 1082 start = avio_tell(pb); 1083 do { 1084 /* save the current offset in case there's nothing to read/skip */ 1085 off = avio_tell(pb); 1086 if (max_search_size && off - start >= max_search_size - ID3v2_HEADER_SIZE) { 1087 avio_seek(pb, off, SEEK_SET); 1088 break; 1089 } 1090 1091 ret = ffio_ensure_seekback(pb, ID3v2_HEADER_SIZE); 1092 if (ret >= 0) 1093 ret = avio_read(pb, buf, ID3v2_HEADER_SIZE); 1094 if (ret != ID3v2_HEADER_SIZE) { 1095 avio_seek(pb, off, SEEK_SET); 1096 break; 1097 } 1098 found_header = ff_id3v2_match(buf, magic); 1099 if (found_header) { 1100 /* parse ID3v2 header */ 1101 len = ((buf[6] & 0x7f) << 21) | 1102 ((buf[7] & 0x7f) << 14) | 1103 ((buf[8] & 0x7f) << 7) | 1104 (buf[9] & 0x7f); 1105 id3v2_parse(pb, metadata, s, len, buf[3], buf[5], extra_meta); 1106 } else { 1107 avio_seek(pb, off, SEEK_SET); 1108 } 1109 } while (found_header); 1110 ff_metadata_conv(metadata, NULL, ff_id3v2_34_metadata_conv); 1111 ff_metadata_conv(metadata, NULL, id3v2_2_metadata_conv); 1112 ff_metadata_conv(metadata, NULL, ff_id3v2_4_metadata_conv); 1113 merge_date(metadata); 1114} 1115 1116void ff_id3v2_read_dict(AVIOContext *pb, AVDictionary **metadata, 1117 const char *magic, ID3v2ExtraMeta **extra_meta) 1118{ 1119 id3v2_read_internal(pb, metadata, NULL, magic, extra_meta, 0); 1120} 1121 1122void ff_id3v2_read(AVFormatContext *s, const char *magic, 1123 ID3v2ExtraMeta **extra_meta, unsigned int max_search_size) 1124{ 1125 id3v2_read_internal(s->pb, &s->metadata, s, magic, extra_meta, max_search_size); 1126} 1127 1128void ff_id3v2_free_extra_meta(ID3v2ExtraMeta **extra_meta) 1129{ 1130 ID3v2ExtraMeta *current = *extra_meta, *next; 1131 const ID3v2EMFunc *extra_func; 1132 1133 while (current) { 1134 if ((extra_func = get_extra_meta_func(current->tag, 1))) 1135 extra_func->free(current->data); 1136 next = current->next; 1137 av_freep(¤t); 1138 current = next; 1139 } 1140 1141 *extra_meta = NULL; 1142} 1143 1144int ff_id3v2_parse_apic(AVFormatContext *s, ID3v2ExtraMeta **extra_meta) 1145{ 1146 ID3v2ExtraMeta *cur; 1147 1148 for (cur = *extra_meta; cur; cur = cur->next) { 1149 ID3v2ExtraMetaAPIC *apic; 1150 AVStream *st; 1151 1152 if (strcmp(cur->tag, "APIC")) 1153 continue; 1154 apic = cur->data; 1155 1156 if (!(st = avformat_new_stream(s, NULL))) 1157 return AVERROR(ENOMEM); 1158 1159 st->disposition |= AV_DISPOSITION_ATTACHED_PIC; 1160 st->codecpar->codec_type = AVMEDIA_TYPE_VIDEO; 1161 st->codecpar->codec_id = apic->id; 1162 1163 if (AV_RB64(apic->buf->data) == PNGSIG) 1164 st->codecpar->codec_id = AV_CODEC_ID_PNG; 1165 1166 if (apic->description[0]) 1167 av_dict_set(&st->metadata, "title", apic->description, 0); 1168 1169 av_dict_set(&st->metadata, "comment", apic->type, 0); 1170 1171 av_init_packet(&st->attached_pic); 1172 st->attached_pic.buf = apic->buf; 1173 st->attached_pic.data = apic->buf->data; 1174 st->attached_pic.size = apic->buf->size - AV_INPUT_BUFFER_PADDING_SIZE; 1175 st->attached_pic.stream_index = st->index; 1176 st->attached_pic.flags |= AV_PKT_FLAG_KEY; 1177 1178 apic->buf = NULL; 1179 } 1180 1181 return 0; 1182} 1183 1184int ff_id3v2_parse_chapters(AVFormatContext *s, ID3v2ExtraMeta **extra_meta) 1185{ 1186 int ret = 0; 1187 ID3v2ExtraMeta *cur; 1188 AVRational time_base = {1, 1000}; 1189 ID3v2ExtraMetaCHAP **chapters = NULL; 1190 int num_chapters = 0; 1191 int i; 1192 1193 // since extra_meta is a linked list where elements are prepended, 1194 // we need to reverse the order of chapters 1195 for (cur = *extra_meta; cur; cur = cur->next) { 1196 ID3v2ExtraMetaCHAP *chap; 1197 1198 if (strcmp(cur->tag, "CHAP")) 1199 continue; 1200 chap = cur->data; 1201 1202 if ((ret = av_dynarray_add_nofree(&chapters, &num_chapters, chap)) < 0) 1203 goto end; 1204 } 1205 1206 for (i = 0; i < (num_chapters / 2); i++) { 1207 ID3v2ExtraMetaCHAP *right; 1208 int right_index; 1209 1210 right_index = (num_chapters - 1) - i; 1211 right = chapters[right_index]; 1212 1213 chapters[right_index] = chapters[i]; 1214 chapters[i] = right; 1215 } 1216 1217 for (i = 0; i < num_chapters; i++) { 1218 ID3v2ExtraMetaCHAP *chap; 1219 AVChapter *chapter; 1220 1221 chap = chapters[i]; 1222 chapter = avpriv_new_chapter(s, i, time_base, chap->start, chap->end, chap->element_id); 1223 if (!chapter) 1224 continue; 1225 1226 if ((ret = av_dict_copy(&chapter->metadata, chap->meta, 0)) < 0) 1227 goto end; 1228 } 1229 1230end: 1231 av_freep(&chapters); 1232 return ret; 1233} 1234 1235int ff_id3v2_parse_priv_dict(AVDictionary **metadata, ID3v2ExtraMeta **extra_meta) 1236{ 1237 ID3v2ExtraMeta *cur; 1238 int dict_flags = AV_DICT_DONT_OVERWRITE | AV_DICT_DONT_STRDUP_KEY | AV_DICT_DONT_STRDUP_VAL; 1239 1240 for (cur = *extra_meta; cur; cur = cur->next) { 1241 if (!strcmp(cur->tag, "PRIV")) { 1242 ID3v2ExtraMetaPRIV *priv = cur->data; 1243 AVBPrint bprint; 1244 char *escaped, *key; 1245 int i, ret; 1246 1247 if ((key = av_asprintf(ID3v2_PRIV_METADATA_PREFIX "%s", priv->owner)) == NULL) { 1248 return AVERROR(ENOMEM); 1249 } 1250 1251 av_bprint_init(&bprint, priv->datasize + 1, AV_BPRINT_SIZE_UNLIMITED); 1252 1253 for (i = 0; i < priv->datasize; i++) { 1254 if (priv->data[i] < 32 || priv->data[i] > 126 || priv->data[i] == '\\') { 1255 av_bprintf(&bprint, "\\x%02x", priv->data[i]); 1256 } else { 1257 av_bprint_chars(&bprint, priv->data[i], 1); 1258 } 1259 } 1260 1261 if ((ret = av_bprint_finalize(&bprint, &escaped)) < 0) { 1262 av_free(key); 1263 return ret; 1264 } 1265 1266 if ((ret = av_dict_set(metadata, key, escaped, dict_flags)) < 0) { 1267 return ret; 1268 } 1269 } 1270 } 1271 1272 return 0; 1273} 1274 1275int ff_id3v2_parse_priv(AVFormatContext *s, ID3v2ExtraMeta **extra_meta) 1276{ 1277 return ff_id3v2_parse_priv_dict(&s->metadata, extra_meta); 1278}