libavformat/id3v2.c

   1 /*
   2  * ID3v2 header parser
   3  * Copyright (c) 2003 Fabrice Bellard
   4  *
   5  * This file is part of Libav.
   6  *
   7  * Libav is free software; you can redistribute it and/or
   8  * modify it under the terms of the GNU Lesser General Public
   9  * License as published by the Free Software Foundation; either
  10  * version 2.1 of the License, or (at your option) any later version.
  11  *
  12  * Libav is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  15  * Lesser General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU Lesser General Public
  18  * License along with Libav; if not, write to the Free Software
  19  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  20  */
  21
  22 #include "id3v2.h"
  23 #include "id3v1.h"
  24 #include "libavutil/avstring.h"
  25 #include "libavutil/intreadwrite.h"
  26 #include "libavutil/dict.h"
  27 #include "avio_internal.h"
  28
  29 const AVMetadataConv ff_id3v2_34_metadata_conv[] = {
  30     { "TALB", "album"},
  31     { "TCOM", "composer"},
  32     { "TCON", "genre"},
  33     { "TCOP", "copyright"},
  34     { "TENC", "encoded_by"},
  35     { "TIT2", "title"},
  36     { "TLAN", "language"},
  37     { "TPE1", "artist"},
  38     { "TPE2", "album_artist"},
  39     { "TPE3", "performer"},
  40     { "TPOS", "disc"},
  41     { "TPUB", "publisher"},
  42     { "TRCK", "track"},
  43     { "TSSE", "encoder"},
  44     { 0 }
  45 };
  46
  47 const AVMetadataConv ff_id3v2_4_metadata_conv[] = {
  48     { "TDRL", "date"},
  49     { "TDRC", "date"},
  50     { "TDEN", "creation_time"},
  51     { "TSOA", "album-sort"},
  52     { "TSOP", "artist-sort"},
  53     { "TSOT", "title-sort"},
  54     { 0 }
  55 };
  56
  57 const AVMetadataConv ff_id3v2_2_metadata_conv[] = {
  58     { "TAL",  "album"},
  59     { "TCO",  "genre"},
  60     { "TT2",  "title"},
  61     { "TEN",  "encoded_by"},
  62     { "TP1",  "artist"},
  63     { "TP2",  "album_artist"},
  64     { "TP3",  "performer"},
  65     { "TRK",  "track"},
  66     { 0 }
  67 };
  68
  69
  70 const char ff_id3v2_tags[][4] = {
  71    "TALB", "TBPM", "TCOM", "TCON", "TCOP", "TDLY", "TENC", "TEXT",
  72    "TFLT", "TIT1", "TIT2", "TIT3", "TKEY", "TLAN", "TLEN", "TMED",
  73    "TOAL", "TOFN", "TOLY", "TOPE", "TOWN", "TPE1", "TPE2", "TPE3",
  74    "TPE4", "TPOS", "TPUB", "TRCK", "TRSN", "TRSO", "TSRC", "TSSE",
  75    { 0 },
  76 };
  77
  78 const char ff_id3v2_4_tags[][4] = {
  79    "TDEN", "TDOR", "TDRC", "TDRL", "TDTG", "TIPL", "TMCL", "TMOO",
  80    "TPRO", "TSOA", "TSOP", "TSOT", "TSST",
  81    { 0 },
  82 };
  83
  84 const char ff_id3v2_3_tags[][4] = {
  85    "TDAT", "TIME", "TORY", "TRDA", "TSIZ", "TYER",
  86    { 0 },
  87 };
  88
  89 int ff_id3v2_match(const uint8_t *buf, const char * magic)
  90 {
  91     return  buf[0]         == magic[0] &&
  92             buf[1]         == magic[1] &&
  93             buf[2]         == magic[2] &&
  94             buf[3]         != 0xff &&
  95             buf[4]         != 0xff &&
  96            (buf[6] & 0x80) ==    0 &&
  97            (buf[7] & 0x80) ==    0 &&
  98            (buf[8] & 0x80) ==    0 &&
  99            (buf[9] & 0x80) ==    0;
 100 }
 101
 102 int ff_id3v2_tag_len(const uint8_t * buf)
 103 {
 104     int len = ((buf[6] & 0x7f) << 21) +
 105               ((buf[7] & 0x7f) << 14) +
 106               ((buf[8] & 0x7f) << 7) +
 107                (buf[9] & 0x7f) +
 108               ID3v2_HEADER_SIZE;
 109     if (buf[5] & 0x10)
 110         len += ID3v2_HEADER_SIZE;
 111     return len;
 112 }
 113
 114 static unsigned int get_size(AVIOContext *s, int len)
 115 {
 116     int v = 0;
 117     while (len--)
 118         v = (v << 7) + (avio_r8(s) & 0x7F);
 119     return v;
 120 }
 121
 122 /**
 123  * Free GEOB type extra metadata.
 124  */
 125 static void free_geobtag(ID3v2ExtraMetaGEOB *geob)
 126 {
 127     av_free(geob->mime_type);
 128     av_free(geob->file_name);
 129     av_free(geob->description);
 130     av_free(geob->data);
 131     av_free(geob);
 132 }
 133
 134 /**
 135  * Decode characters to UTF-8 according to encoding type. The decoded buffer is
 136  * always null terminated. Stop reading when either *maxread bytes are read from
 137  * pb or U+0000 character is found.
 138  *
 139  * @param dst Pointer where the address of the buffer with the decoded bytes is
 140  * stored. Buffer must be freed by caller.
 141  * @param maxread Pointer to maximum number of characters to read from the
 142  * AVIOContext. After execution the value is decremented by the number of bytes
 143  * actually read.
 144  * @returns 0 if no error occured, dst is uninitialized on error
 145  */
 146 static int decode_str(AVFormatContext *s, AVIOContext *pb, int encoding,
 147                       uint8_t **dst, int *maxread)
 148 {
 149     int ret;
 150     uint8_t tmp;
 151     uint32_t ch = 1;
 152     int left = *maxread;
 153     unsigned int (*get)(AVIOContext*) = avio_rb16;
 154     AVIOContext *dynbuf;
 155
 156     if ((ret = avio_open_dyn_buf(&dynbuf)) < 0) {
 157         av_log(s, AV_LOG_ERROR, "Error opening memory stream\n");
 158         return ret;
 159     }
 160
 161     switch (encoding) {
 162
 163     case ID3v2_ENCODING_ISO8859:
 164         while (left && ch) {
 165             ch = avio_r8(pb);
 166             PUT_UTF8(ch, tmp, avio_w8(dynbuf, tmp);)
 167             left--;
 168         }
 169         break;
 170
 171     case ID3v2_ENCODING_UTF16BOM:
 172         if ((left -= 2) < 0) {
 173             av_log(s, AV_LOG_ERROR, "Cannot read BOM value, input too short\n");
 174             avio_close_dyn_buf(dynbuf, dst);
 175             av_freep(dst);
 176             return AVERROR_INVALIDDATA;
 177         }
 178         switch (avio_rb16(pb)) {
 179         case 0xfffe:
 180             get = avio_rl16;
 181         case 0xfeff:
 182             break;
 183         default:
 184             av_log(s, AV_LOG_ERROR, "Incorrect BOM value\n");
 185             avio_close_dyn_buf(dynbuf, dst);
 186             av_freep(dst);
 187             *maxread = left;
 188             return AVERROR_INVALIDDATA;
 189         }
 190         // fall-through
 191
 192     case ID3v2_ENCODING_UTF16BE:
 193         while ((left > 1) && ch) {
 194             GET_UTF16(ch, ((left -= 2) >= 0 ? get(pb) : 0), break;)
 195             PUT_UTF8(ch, tmp, avio_w8(dynbuf, tmp);)
 196         }
 197         if (left < 0)
 198             left += 2; /* did not read last char from pb */
 199         break;
 200
 201     case ID3v2_ENCODING_UTF8:
 202         while (left && ch) {
 203             ch = avio_r8(pb);
 204             avio_w8(dynbuf, ch);
 205             left--;
 206         }
 207         break;
 208     default:
 209         av_log(s, AV_LOG_WARNING, "Unknown encoding\n");
 210     }
 211
 212     if (ch)
 213         avio_w8(dynbuf, 0);
 214
 215     avio_close_dyn_buf(dynbuf, dst);
 216     *maxread = left;
 217
 218     return 0;
 219 }
 220
 221 /**
 222  * Parse a text tag.
 223  */
 224 static void read_ttag(AVFormatContext *s, AVIOContext *pb, int taglen, const char *key)
 225 {
 226     uint8_t *dst;
 227     int encoding, dict_flags = AV_DICT_DONT_OVERWRITE;
 228     unsigned genre;
 229
 230     if (taglen < 1)
 231         return;
 232
 233     encoding = avio_r8(pb);
 234     taglen--; /* account for encoding type byte */
 235
 236     if (decode_str(s, pb, encoding, &dst, &taglen) < 0) {
 237         av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", key);
 238         return;
 239     }
 240
 241     if (!(strcmp(key, "TCON") && strcmp(key, "TCO"))
 242         && (sscanf(dst, "(%d)", &genre) == 1 || sscanf(dst, "%d", &genre) == 1)
 243         && genre <= ID3v1_GENRE_MAX) {
 244         av_freep(&dst);
 245         dst = ff_id3v1_genre_str[genre];
 246     } else if (!(strcmp(key, "TXXX") && strcmp(key, "TXX"))) {
 247         /* dst now contains the key, need to get value */
 248         key = dst;
 249         if (decode_str(s, pb, encoding, &dst, &taglen) < 0) {
 250             av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", key);
 251             av_freep(&key);
 252             return;
 253         }
 254         dict_flags |= AV_DICT_DONT_STRDUP_VAL | AV_DICT_DONT_STRDUP_KEY;
 255     }
 256     else if (*dst)
 257         dict_flags |= AV_DICT_DONT_STRDUP_VAL;
 258
 259     if (dst)
 260         av_dict_set(&s->metadata, key, dst, dict_flags);
 261 }
 262
 263 /**
 264  * Parse GEOB tag into a ID3v2ExtraMetaGEOB struct.
 265  */
 266 static void read_geobtag(AVFormatContext *s, AVIOContext *pb, int taglen, char *tag, ID3v2ExtraMeta **extra_meta)
 267 {
 268     ID3v2ExtraMetaGEOB *geob_data = NULL;
 269     ID3v2ExtraMeta *new_extra = NULL;
 270     char encoding;
 271     unsigned int len;
 272
 273     if (taglen < 1)
 274         return;
 275
 276     geob_data = av_mallocz(sizeof(ID3v2ExtraMetaGEOB));
 277     if (!geob_data) {
 278         av_log(s, AV_LOG_ERROR, "Failed to alloc %zu bytes\n", sizeof(ID3v2ExtraMetaGEOB));
 279         return;
 280     }
 281
 282     new_extra = av_mallocz(sizeof(ID3v2ExtraMeta));
 283     if (!new_extra) {
 284         av_log(s, AV_LOG_ERROR, "Failed to alloc %zu bytes\n", sizeof(ID3v2ExtraMeta));
 285         goto fail;
 286     }
 287
 288     /* read encoding type byte */
 289     encoding = avio_r8(pb);
 290     taglen--;
 291
 292     /* read MIME type (always ISO-8859) */
 293     if (decode_str(s, pb, ID3v2_ENCODING_ISO8859, &geob_data->mime_type, &taglen) < 0
 294         || taglen <= 0)
 295         goto fail;
 296
 297     /* read file name */
 298     if (decode_str(s, pb, encoding, &geob_data->file_name, &taglen) < 0
 299         || taglen <= 0)
 300         goto fail;
 301
 302     /* read content description */
 303     if (decode_str(s, pb, encoding, &geob_data->description, &taglen) < 0
 304         || taglen < 0)
 305         goto fail;
 306
 307     if (taglen) {
 308         /* save encapsulated binary data */
 309         geob_data->data = av_malloc(taglen);
 310         if (!geob_data->data) {
 311             av_log(s, AV_LOG_ERROR, "Failed to alloc %d bytes\n", taglen);
 312             goto fail;
 313         }
 314         if ((len = avio_read(pb, geob_data->data, taglen)) < taglen)
 315             av_log(s, AV_LOG_WARNING, "Error reading GEOB frame, data truncated.\n");
 316         geob_data->datasize = len;
 317     } else {
 318         geob_data->data = NULL;
 319         geob_data->datasize = 0;
 320     }
 321
 322     /* add data to the list */
 323     new_extra->tag = "GEOB";
 324     new_extra->data = geob_data;
 325     new_extra->next = *extra_meta;
 326     *extra_meta = new_extra;
 327
 328     return;
 329
 330 fail:
 331     av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", tag);
 332     free_geobtag(geob_data);
 333     av_free(new_extra);
 334     return;
 335 }
 336
 337 static int is_number(const char *str)
 338 {
 339     while (*str >= '0' && *str <= '9') str++;
 340     return !*str;
 341 }
 342
 343 static AVDictionaryEntry* get_date_tag(AVDictionary *m, const char *tag)
 344 {
 345     AVDictionaryEntry *t;
 346     if ((t = av_dict_get(m, tag, NULL, AV_DICT_MATCH_CASE)) &&
 347         strlen(t->value) == 4 && is_number(t->value))
 348         return t;
 349     return NULL;
 350 }
 351
 352 static void merge_date(AVDictionary **m)
 353 {
 354     AVDictionaryEntry *t;
 355     char date[17] = {0};      // YYYY-MM-DD hh:mm
 356
 357     if (!(t = get_date_tag(*m, "TYER")) &&
 358         !(t = get_date_tag(*m, "TYE")))
 359         return;
 360     av_strlcpy(date, t->value, 5);
 361     av_dict_set(m, "TYER", NULL, 0);
 362     av_dict_set(m, "TYE",  NULL, 0);
 363
 364     if (!(t = get_date_tag(*m, "TDAT")) &&
 365         !(t = get_date_tag(*m, "TDA")))
 366         goto finish;
 367     snprintf(date + 4, sizeof(date) - 4, "-%.2s-%.2s", t->value + 2, t->value);
 368     av_dict_set(m, "TDAT", NULL, 0);
 369     av_dict_set(m, "TDA",  NULL, 0);
 370
 371     if (!(t = get_date_tag(*m, "TIME")) &&
 372         !(t = get_date_tag(*m, "TIM")))
 373         goto finish;
 374     snprintf(date + 10, sizeof(date) - 10, " %.2s:%.2s", t->value, t->value + 2);
 375     av_dict_set(m, "TIME", NULL, 0);
 376     av_dict_set(m, "TIM",  NULL, 0);
 377
 378 finish:
 379     if (date[0])
 380         av_dict_set(m, "date", date, 0);
 381 }
 382
 383 const ID3v2EMFunc ff_id3v2_extra_meta_funcs[] = {
 384     { "GEO", "GEOB", read_geobtag, free_geobtag },
 385     { NULL }
 386 };
 387
 388 /**
 389  * Get the corresponding ID3v2EMFunc struct for a tag.
 390  * @param isv34 Determines if v2.2 or v2.3/4 strings are used
 391  * @return A pointer to the ID3v2EMFunc struct if found, NULL otherwise.
 392  */
 393 static const ID3v2EMFunc *get_extra_meta_func(const char *tag, int isv34)
 394 {
 395     int i = 0;
 396     while (ff_id3v2_extra_meta_funcs[i].tag3) {
 397         if (!memcmp(tag,
 398                     (isv34 ?
 399                         ff_id3v2_extra_meta_funcs[i].tag4 :
 400                         ff_id3v2_extra_meta_funcs[i].tag3),
 401                     (isv34 ? 4 : 3)))
 402             return &ff_id3v2_extra_meta_funcs[i];
 403         i++;
 404     }
 405     return NULL;
 406 }
 407
 408 static void ff_id3v2_parse(AVFormatContext *s, int len, uint8_t version, uint8_t flags, ID3v2ExtraMeta **extra_meta)
 409 {
 410     int isv34, tlen, unsync;
 411     char tag[5];
 412     int64_t next, end = avio_tell(s->pb) + len;
 413     int taghdrlen;
 414     const char *reason = NULL;
 415     AVIOContext pb;
 416     AVIOContext *pbx;
 417     unsigned char *buffer = NULL;
 418     int buffer_size = 0;
 419     const ID3v2EMFunc *extra_func;
 420
 421     switch (version) {
 422     case 2:
 423         if (flags & 0x40) {
 424             reason = "compression";
 425             goto error;
 426         }
 427         isv34 = 0;
 428         taghdrlen = 6;
 429         break;
 430
 431     case 3:
 432     case 4:
 433         isv34 = 1;
 434         taghdrlen = 10;
 435         break;
 436
 437     default:
 438         reason = "version";
 439         goto error;
 440     }
 441
 442     unsync = flags & 0x80;
 443
 444     if (isv34 && flags & 0x40) /* Extended header present, just skip over it */
 445         avio_skip(s->pb, get_size(s->pb, 4));
 446
 447     while (len >= taghdrlen) {
 448         unsigned int tflags = 0;
 449         int tunsync = 0;
 450
 451         if (isv34) {
 452             avio_read(s->pb, tag, 4);
 453             tag[4] = 0;
 454             if(version==3){
 455                 tlen = avio_rb32(s->pb);
 456             }else
 457                 tlen = get_size(s->pb, 4);
 458             tflags = avio_rb16(s->pb);
 459             tunsync = tflags & ID3v2_FLAG_UNSYNCH;
 460         } else {
 461             avio_read(s->pb, tag, 3);
 462             tag[3] = 0;
 463             tlen = avio_rb24(s->pb);
 464         }
 465         if (tlen < 0 || tlen > len - taghdrlen) {
 466             av_log(s, AV_LOG_WARNING, "Invalid size in frame %s, skipping the rest of tag.\n", tag);
 467             break;
 468         }
 469         len -= taghdrlen + tlen;
 470         next = avio_tell(s->pb) + tlen;
 471
 472         if (!tlen) {
 473             if (tag[0])
 474                 av_log(s, AV_LOG_DEBUG, "Invalid empty frame %s, skipping.\n", tag);
 475             continue;
 476         }
 477
 478         if (tflags & ID3v2_FLAG_DATALEN) {
 479             avio_rb32(s->pb);
 480             tlen -= 4;
 481         }
 482
 483         if (tflags & (ID3v2_FLAG_ENCRYPTION | ID3v2_FLAG_COMPRESSION)) {
 484             av_log(s, AV_LOG_WARNING, "Skipping encrypted/compressed ID3v2 frame %s.\n", tag);
 485             avio_skip(s->pb, tlen);
 486         /* check for text tag or supported special meta tag */
 487         } else if (tag[0] == 'T' || (extra_meta && (extra_func = get_extra_meta_func(tag, isv34)))) {
 488             if (unsync || tunsync) {
 489                 int i, j;
 490                 av_fast_malloc(&buffer, &buffer_size, tlen);
 491                 if (!buffer) {
 492                     av_log(s, AV_LOG_ERROR, "Failed to alloc %d bytes\n", tlen);
 493                     goto seek;
 494                 }
 495                 for (i = 0, j = 0; i < tlen; i++, j++) {
 496                     buffer[j] = avio_r8(s->pb);
 497                     if (j > 0 && !buffer[j] && buffer[j - 1] == 0xff) {
 498                         /* Unsynchronised byte, skip it */
 499                         j--;
 500                     }
 501                 }
 502                 ffio_init_context(&pb, buffer, j, 0, NULL, NULL, NULL, NULL);
 503                 tlen = j;
 504                 pbx = &pb; // read from sync buffer
 505             } else {
 506                 pbx = s->pb; // read straight from input
 507             }
 508             if (tag[0] == 'T')
 509                 /* parse text tag */
 510                 read_ttag(s, pbx, tlen, tag);
 511             else
 512                 /* parse special meta tag */
 513                 extra_func->read(s, pbx, tlen, tag, extra_meta);
 514         }
 515         else if (!tag[0]) {
 516             if (tag[1])
 517                 av_log(s, AV_LOG_WARNING, "invalid frame id, assuming padding");
 518             avio_skip(s->pb, tlen);
 519             break;
 520         }
 521         /* Skip to end of tag */
 522 seek:
 523         avio_seek(s->pb, next, SEEK_SET);
 524     }
 525
 526     if (version == 4 && flags & 0x10) /* Footer preset, always 10 bytes, skip over it */
 527         end += 10;
 528
 529   error:
 530     if (reason)
 531         av_log(s, AV_LOG_INFO, "ID3v2.%d tag skipped, cannot handle %s\n", version, reason);
 532     avio_seek(s->pb, end, SEEK_SET);
 533     av_free(buffer);
 534     return;
 535 }
 536
 537 void ff_id3v2_read_all(AVFormatContext *s, const char *magic, ID3v2ExtraMeta **extra_meta)
 538 {
 539     int len, ret;
 540     uint8_t buf[ID3v2_HEADER_SIZE];
 541     int     found_header;
 542     int64_t off;
 543
 544     do {
 545         /* save the current offset in case there's nothing to read/skip */
 546         off = avio_tell(s->pb);
 547         ret = avio_read(s->pb, buf, ID3v2_HEADER_SIZE);
 548         if (ret != ID3v2_HEADER_SIZE)
 549             break;
 550             found_header = ff_id3v2_match(buf, magic);
 551             if (found_header) {
 552             /* parse ID3v2 header */
 553             len = ((buf[6] & 0x7f) << 21) |
 554                   ((buf[7] & 0x7f) << 14) |
 555                   ((buf[8] & 0x7f) << 7) |
 556                    (buf[9] & 0x7f);
 557             ff_id3v2_parse(s, len, buf[3], buf[5], extra_meta);
 558         } else {
 559             avio_seek(s->pb, off, SEEK_SET);
 560         }
 561     } while (found_header);
 562     ff_metadata_conv(&s->metadata, NULL, ff_id3v2_34_metadata_conv);
 563     ff_metadata_conv(&s->metadata, NULL, ff_id3v2_2_metadata_conv);
 564     ff_metadata_conv(&s->metadata, NULL, ff_id3v2_4_metadata_conv);
 565     merge_date(&s->metadata);
 566 }
 567
 568 void ff_id3v2_read(AVFormatContext *s, const char *magic)
 569 {
 570     ff_id3v2_read_all(s, magic, NULL);
 571 }
 572
 573 void ff_id3v2_free_extra_meta(ID3v2ExtraMeta **extra_meta)
 574 {
 575     ID3v2ExtraMeta *current = *extra_meta, *next;
 576     const ID3v2EMFunc *extra_func;
 577
 578     while (current) {
 579         if ((extra_func = get_extra_meta_func(current->tag, 1)))
 580             extra_func->free(current->data);
 581         next = current->next;
 582         av_freep(&current);
 583         current = next;
 584     }
 585 }