FFmpeg
 All Data Structures Namespaces Files Functions Variables Typedefs Enumerations Enumerator Macros Groups Pages
id3v2.c
Go to the documentation of this 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/dict.h"
37 #include "libavutil/intreadwrite.h"
38 #include "avio_internal.h"
39 #include "internal.h"
40 #include "id3v1.h"
41 #include "id3v2.h"
42 
44  { "TALB", "album" },
45  { "TCOM", "composer" },
46  { "TCON", "genre" },
47  { "TCOP", "copyright" },
48  { "TENC", "encoded_by" },
49  { "TIT2", "title" },
50  { "TLAN", "language" },
51  { "TPE1", "artist" },
52  { "TPE2", "album_artist" },
53  { "TPE3", "performer" },
54  { "TPOS", "disc" },
55  { "TPUB", "publisher" },
56  { "TRCK", "track" },
57  { "TSSE", "encoder" },
58  { "USLT", "lyrics" },
59  { 0 }
60 };
61 
63  { "TCMP", "compilation" },
64  { "TDRC", "date" },
65  { "TDRL", "date" },
66  { "TDEN", "creation_time" },
67  { "TSOA", "album-sort" },
68  { "TSOP", "artist-sort" },
69  { "TSOT", "title-sort" },
70  { 0 }
71 };
72 
74  { "TAL", "album" },
75  { "TCO", "genre" },
76  { "TCP", "compilation" },
77  { "TT2", "title" },
78  { "TEN", "encoded_by" },
79  { "TP1", "artist" },
80  { "TP2", "album_artist" },
81  { "TP3", "performer" },
82  { "TRK", "track" },
83  { 0 }
84 };
85 
86 const char ff_id3v2_tags[][4] = {
87  "TALB", "TBPM", "TCOM", "TCON", "TCOP", "TDLY", "TENC", "TEXT",
88  "TFLT", "TIT1", "TIT2", "TIT3", "TKEY", "TLAN", "TLEN", "TMED",
89  "TOAL", "TOFN", "TOLY", "TOPE", "TOWN", "TPE1", "TPE2", "TPE3",
90  "TPE4", "TPOS", "TPUB", "TRCK", "TRSN", "TRSO", "TSRC", "TSSE",
91  { 0 },
92 };
93 
94 const char ff_id3v2_4_tags[][4] = {
95  "TDEN", "TDOR", "TDRC", "TDRL", "TDTG", "TIPL", "TMCL", "TMOO",
96  "TPRO", "TSOA", "TSOP", "TSOT", "TSST",
97  { 0 },
98 };
99 
100 const char ff_id3v2_3_tags[][4] = {
101  "TDAT", "TIME", "TORY", "TRDA", "TSIZ", "TYER",
102  { 0 },
103 };
104 
105 const char *ff_id3v2_picture_types[21] = {
106  "Other",
107  "32x32 pixels 'file icon'",
108  "Other file icon",
109  "Cover (front)",
110  "Cover (back)",
111  "Leaflet page",
112  "Media (e.g. label side of CD)",
113  "Lead artist/lead performer/soloist",
114  "Artist/performer",
115  "Conductor",
116  "Band/Orchestra",
117  "Composer",
118  "Lyricist/text writer",
119  "Recording Location",
120  "During recording",
121  "During performance",
122  "Movie/video screen capture",
123  "A bright coloured fish",
124  "Illustration",
125  "Band/artist logotype",
126  "Publisher/Studio logotype",
127 };
128 
130  { "image/gif", AV_CODEC_ID_GIF },
131  { "image/jpeg", AV_CODEC_ID_MJPEG },
132  { "image/jpg", AV_CODEC_ID_MJPEG },
133  { "image/png", AV_CODEC_ID_PNG },
134  { "image/tiff", AV_CODEC_ID_TIFF },
135  { "image/bmp", AV_CODEC_ID_BMP },
136  { "JPG", AV_CODEC_ID_MJPEG }, /* ID3v2.2 */
137  { "PNG", AV_CODEC_ID_PNG }, /* ID3v2.2 */
138  { "", AV_CODEC_ID_NONE },
139 };
140 
141 int ff_id3v2_match(const uint8_t *buf, const char *magic)
142 {
143  return buf[0] == magic[0] &&
144  buf[1] == magic[1] &&
145  buf[2] == magic[2] &&
146  buf[3] != 0xff &&
147  buf[4] != 0xff &&
148  (buf[6] & 0x80) == 0 &&
149  (buf[7] & 0x80) == 0 &&
150  (buf[8] & 0x80) == 0 &&
151  (buf[9] & 0x80) == 0;
152 }
153 
155 {
156  int len = ((buf[6] & 0x7f) << 21) +
157  ((buf[7] & 0x7f) << 14) +
158  ((buf[8] & 0x7f) << 7) +
159  (buf[9] & 0x7f) +
161  if (buf[5] & 0x10)
162  len += ID3v2_HEADER_SIZE;
163  return len;
164 }
165 
166 static unsigned int get_size(AVIOContext *s, int len)
167 {
168  int v = 0;
169  while (len--)
170  v = (v << 7) + (avio_r8(s) & 0x7F);
171  return v;
172 }
173 
174 static unsigned int size_to_syncsafe(unsigned int size)
175 {
176  return (((size) & (0x7f << 0)) >> 0) +
177  (((size) & (0x7f << 8)) >> 1) +
178  (((size) & (0x7f << 16)) >> 2) +
179  (((size) & (0x7f << 24)) >> 3);
180 }
181 
182 /* No real verification, only check that the tag consists of
183  * a combination of capital alpha-numerical characters */
184 static int is_tag(const char *buf, unsigned int len)
185 {
186  if (!len)
187  return 0;
188 
189  while (len--)
190  if ((buf[len] < 'A' ||
191  buf[len] > 'Z') &&
192  (buf[len] < '0' ||
193  buf[len] > '9'))
194  return 0;
195 
196  return 1;
197 }
198 
199 /**
200  * Return 1 if the tag of length len at the given offset is valid, 0 if not, -1 on error
201  */
202 static int check_tag(AVIOContext *s, int offset, unsigned int len)
203 {
204  char tag[4];
205 
206  if (len > 4 ||
207  avio_seek(s, offset, SEEK_SET) < 0 ||
208  avio_read(s, tag, len) < len)
209  return -1;
210  else if (!AV_RB32(tag) || is_tag(tag, len))
211  return 1;
212 
213  return 0;
214 }
215 
216 /**
217  * Free GEOB type extra metadata.
218  */
219 static void free_geobtag(void *obj)
220 {
221  ID3v2ExtraMetaGEOB *geob = obj;
222  av_freep(&geob->mime_type);
223  av_freep(&geob->file_name);
224  av_freep(&geob->description);
225  av_freep(&geob->data);
226  av_free(geob);
227 }
228 
229 /**
230  * Decode characters to UTF-8 according to encoding type. The decoded buffer is
231  * always null terminated. Stop reading when either *maxread bytes are read from
232  * pb or U+0000 character is found.
233  *
234  * @param dst Pointer where the address of the buffer with the decoded bytes is
235  * stored. Buffer must be freed by caller.
236  * @param maxread Pointer to maximum number of characters to read from the
237  * AVIOContext. After execution the value is decremented by the number of bytes
238  * actually read.
239  * @returns 0 if no error occurred, dst is uninitialized on error
240  */
241 static int decode_str(AVFormatContext *s, AVIOContext *pb, int encoding,
242  uint8_t **dst, int *maxread)
243 {
244  int ret;
245  uint8_t tmp;
246  uint32_t ch = 1;
247  int left = *maxread;
248  unsigned int (*get)(AVIOContext*) = avio_rb16;
249  AVIOContext *dynbuf;
250 
251  if ((ret = avio_open_dyn_buf(&dynbuf)) < 0) {
252  av_log(s, AV_LOG_ERROR, "Error opening memory stream\n");
253  return ret;
254  }
255 
256  switch (encoding) {
258  while (left && ch) {
259  ch = avio_r8(pb);
260  PUT_UTF8(ch, tmp, avio_w8(dynbuf, tmp);)
261  left--;
262  }
263  break;
264 
266  if ((left -= 2) < 0) {
267  av_log(s, AV_LOG_ERROR, "Cannot read BOM value, input too short\n");
268  avio_close_dyn_buf(dynbuf, dst);
269  av_freep(dst);
270  return AVERROR_INVALIDDATA;
271  }
272  switch (avio_rb16(pb)) {
273  case 0xfffe:
274  get = avio_rl16;
275  case 0xfeff:
276  break;
277  default:
278  av_log(s, AV_LOG_ERROR, "Incorrect BOM value\n");
279  avio_close_dyn_buf(dynbuf, dst);
280  av_freep(dst);
281  *maxread = left;
282  return AVERROR_INVALIDDATA;
283  }
284  // fall-through
285 
287  while ((left > 1) && ch) {
288  GET_UTF16(ch, ((left -= 2) >= 0 ? get(pb) : 0), break;)
289  PUT_UTF8(ch, tmp, avio_w8(dynbuf, tmp);)
290  }
291  if (left < 0)
292  left += 2; /* did not read last char from pb */
293  break;
294 
295  case ID3v2_ENCODING_UTF8:
296  while (left && ch) {
297  ch = avio_r8(pb);
298  avio_w8(dynbuf, ch);
299  left--;
300  }
301  break;
302  default:
303  av_log(s, AV_LOG_WARNING, "Unknown encoding\n");
304  }
305 
306  if (ch)
307  avio_w8(dynbuf, 0);
308 
309  avio_close_dyn_buf(dynbuf, dst);
310  *maxread = left;
311 
312  return 0;
313 }
314 
315 /**
316  * Parse a text tag.
317  */
318 static void read_ttag(AVFormatContext *s, AVIOContext *pb, int taglen,
319  AVDictionary **metadata, const char *key)
320 {
321  uint8_t *dst;
322  int encoding, dict_flags = AV_DICT_DONT_OVERWRITE | AV_DICT_DONT_STRDUP_VAL;
323  unsigned genre;
324 
325  if (taglen < 1)
326  return;
327 
328  encoding = avio_r8(pb);
329  taglen--; /* account for encoding type byte */
330 
331  if (decode_str(s, pb, encoding, &dst, &taglen) < 0) {
332  av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", key);
333  return;
334  }
335 
336  if (!(strcmp(key, "TCON") && strcmp(key, "TCO")) &&
337  (sscanf(dst, "(%d)", &genre) == 1 || sscanf(dst, "%d", &genre) == 1) &&
338  genre <= ID3v1_GENRE_MAX) {
339  av_freep(&dst);
340  dst = av_strdup(ff_id3v1_genre_str[genre]);
341  } else if (!(strcmp(key, "TXXX") && strcmp(key, "TXX"))) {
342  /* dst now contains the key, need to get value */
343  key = dst;
344  if (decode_str(s, pb, encoding, &dst, &taglen) < 0) {
345  av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", key);
346  av_freep(&key);
347  return;
348  }
349  dict_flags |= AV_DICT_DONT_STRDUP_KEY;
350  } else if (!*dst)
351  av_freep(&dst);
352 
353  if (dst)
354  av_dict_set(metadata, key, dst, dict_flags);
355 }
356 
357 static void read_uslt(AVFormatContext *s, AVIOContext *pb, int taglen,
358  AVDictionary **metadata)
359 {
360  uint8_t lang[4];
361  uint8_t *descriptor = NULL; // 'Content descriptor'
362  uint8_t *text = NULL;
363  char *key = NULL;
364  int encoding;
365  int ok = 0;
366 
367  if (taglen < 1)
368  goto error;
369 
370  encoding = avio_r8(pb);
371  taglen--;
372 
373  if (avio_read(pb, lang, 3) < 3)
374  goto error;
375  lang[3] = '\0';
376  taglen -= 3;
377 
378  if (decode_str(s, pb, encoding, &descriptor, &taglen) < 0)
379  goto error;
380 
381  if (decode_str(s, pb, encoding, &text, &taglen) < 0)
382  goto error;
383 
384  // FFmpeg does not support hierarchical metadata, so concatenate the keys.
385  key = av_asprintf("lyrics-%s%s%s", descriptor[0] ? (char *)descriptor : "",
386  descriptor[0] ? "-" : "",
387  lang);
388  if (!key)
389  goto error;
390 
391  av_dict_set(metadata, key, text, 0);
392 
393  ok = 1;
394 error:
395  if (!ok)
396  av_log(s, AV_LOG_ERROR, "Error reading lyrics, skipped\n");
397  av_free(descriptor);
398  av_free(text);
399  av_free(key);
400 }
401 
402 /**
403  * Parse GEOB tag into a ID3v2ExtraMetaGEOB struct.
404  */
405 static void read_geobtag(AVFormatContext *s, AVIOContext *pb, int taglen,
406  const char *tag, ID3v2ExtraMeta **extra_meta,
407  int isv34)
408 {
409  ID3v2ExtraMetaGEOB *geob_data = NULL;
410  ID3v2ExtraMeta *new_extra = NULL;
411  char encoding;
412  unsigned int len;
413 
414  if (taglen < 1)
415  return;
416 
417  geob_data = av_mallocz(sizeof(ID3v2ExtraMetaGEOB));
418  if (!geob_data) {
419  av_log(s, AV_LOG_ERROR, "Failed to alloc %"SIZE_SPECIFIER" bytes\n",
420  sizeof(ID3v2ExtraMetaGEOB));
421  return;
422  }
423 
424  new_extra = av_mallocz(sizeof(ID3v2ExtraMeta));
425  if (!new_extra) {
426  av_log(s, AV_LOG_ERROR, "Failed to alloc %"SIZE_SPECIFIER" bytes\n",
427  sizeof(ID3v2ExtraMeta));
428  goto fail;
429  }
430 
431  /* read encoding type byte */
432  encoding = avio_r8(pb);
433  taglen--;
434 
435  /* read MIME type (always ISO-8859) */
436  if (decode_str(s, pb, ID3v2_ENCODING_ISO8859, &geob_data->mime_type,
437  &taglen) < 0 ||
438  taglen <= 0)
439  goto fail;
440 
441  /* read file name */
442  if (decode_str(s, pb, encoding, &geob_data->file_name, &taglen) < 0 ||
443  taglen <= 0)
444  goto fail;
445 
446  /* read content description */
447  if (decode_str(s, pb, encoding, &geob_data->description, &taglen) < 0 ||
448  taglen < 0)
449  goto fail;
450 
451  if (taglen) {
452  /* save encapsulated binary data */
453  geob_data->data = av_malloc(taglen);
454  if (!geob_data->data) {
455  av_log(s, AV_LOG_ERROR, "Failed to alloc %d bytes\n", taglen);
456  goto fail;
457  }
458  if ((len = avio_read(pb, geob_data->data, taglen)) < taglen)
460  "Error reading GEOB frame, data truncated.\n");
461  geob_data->datasize = len;
462  } else {
463  geob_data->data = NULL;
464  geob_data->datasize = 0;
465  }
466 
467  /* add data to the list */
468  new_extra->tag = "GEOB";
469  new_extra->data = geob_data;
470  new_extra->next = *extra_meta;
471  *extra_meta = new_extra;
472 
473  return;
474 
475 fail:
476  av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", tag);
477  free_geobtag(geob_data);
478  av_free(new_extra);
479  return;
480 }
481 
482 static int is_number(const char *str)
483 {
484  while (*str >= '0' && *str <= '9')
485  str++;
486  return !*str;
487 }
488 
490 {
492  if ((t = av_dict_get(m, tag, NULL, AV_DICT_MATCH_CASE)) &&
493  strlen(t->value) == 4 && is_number(t->value))
494  return t;
495  return NULL;
496 }
497 
498 static void merge_date(AVDictionary **m)
499 {
501  char date[17] = { 0 }; // YYYY-MM-DD hh:mm
502 
503  if (!(t = get_date_tag(*m, "TYER")) &&
504  !(t = get_date_tag(*m, "TYE")))
505  return;
506  av_strlcpy(date, t->value, 5);
507  av_dict_set(m, "TYER", NULL, 0);
508  av_dict_set(m, "TYE", NULL, 0);
509 
510  if (!(t = get_date_tag(*m, "TDAT")) &&
511  !(t = get_date_tag(*m, "TDA")))
512  goto finish;
513  snprintf(date + 4, sizeof(date) - 4, "-%.2s-%.2s", t->value + 2, t->value);
514  av_dict_set(m, "TDAT", NULL, 0);
515  av_dict_set(m, "TDA", NULL, 0);
516 
517  if (!(t = get_date_tag(*m, "TIME")) &&
518  !(t = get_date_tag(*m, "TIM")))
519  goto finish;
520  snprintf(date + 10, sizeof(date) - 10,
521  " %.2s:%.2s", t->value, t->value + 2);
522  av_dict_set(m, "TIME", NULL, 0);
523  av_dict_set(m, "TIM", NULL, 0);
524 
525 finish:
526  if (date[0])
527  av_dict_set(m, "date", date, 0);
528 }
529 
530 static void free_apic(void *obj)
531 {
532  ID3v2ExtraMetaAPIC *apic = obj;
533  av_buffer_unref(&apic->buf);
534  av_freep(&apic->description);
535  av_freep(&apic);
536 }
537 
538 static void read_apic(AVFormatContext *s, AVIOContext *pb, int taglen,
539  const char *tag, ID3v2ExtraMeta **extra_meta,
540  int isv34)
541 {
542  int enc, pic_type;
543  char mimetype[64];
544  const CodecMime *mime = ff_id3v2_mime_tags;
545  enum AVCodecID id = AV_CODEC_ID_NONE;
546  ID3v2ExtraMetaAPIC *apic = NULL;
547  ID3v2ExtraMeta *new_extra = NULL;
548  int64_t end = avio_tell(pb) + taglen;
549 
550  if (taglen <= 4 || (!isv34 && taglen <= 6))
551  goto fail;
552 
553  new_extra = av_mallocz(sizeof(*new_extra));
554  apic = av_mallocz(sizeof(*apic));
555  if (!new_extra || !apic)
556  goto fail;
557 
558  enc = avio_r8(pb);
559  taglen--;
560 
561  /* mimetype */
562  if (isv34) {
563  taglen -= avio_get_str(pb, taglen, mimetype, sizeof(mimetype));
564  } else {
565  avio_read(pb, mimetype, 3);
566  mimetype[3] = 0;
567  taglen -= 3;
568  }
569 
570  while (mime->id != AV_CODEC_ID_NONE) {
571  if (!av_strncasecmp(mime->str, mimetype, sizeof(mimetype))) {
572  id = mime->id;
573  break;
574  }
575  mime++;
576  }
577  if (id == AV_CODEC_ID_NONE) {
579  "Unknown attached picture mimetype: %s, skipping.\n", mimetype);
580  goto fail;
581  }
582  apic->id = id;
583 
584  /* picture type */
585  pic_type = avio_r8(pb);
586  taglen--;
587  if (pic_type < 0 || pic_type >= FF_ARRAY_ELEMS(ff_id3v2_picture_types)) {
588  av_log(s, AV_LOG_WARNING, "Unknown attached picture type %d.\n",
589  pic_type);
590  pic_type = 0;
591  }
592  apic->type = ff_id3v2_picture_types[pic_type];
593 
594  /* description and picture data */
595  if (decode_str(s, pb, enc, &apic->description, &taglen) < 0) {
596  av_log(s, AV_LOG_ERROR,
597  "Error decoding attached picture description.\n");
598  goto fail;
599  }
600 
602  if (!apic->buf || !taglen || avio_read(pb, apic->buf->data, taglen) != taglen)
603  goto fail;
604  memset(apic->buf->data + taglen, 0, FF_INPUT_BUFFER_PADDING_SIZE);
605 
606  new_extra->tag = "APIC";
607  new_extra->data = apic;
608  new_extra->next = *extra_meta;
609  *extra_meta = new_extra;
610 
611  return;
612 
613 fail:
614  if (apic)
615  free_apic(apic);
616  av_freep(&new_extra);
617  avio_seek(pb, end, SEEK_SET);
618 }
619 
620 static void read_chapter(AVFormatContext *s, AVIOContext *pb, int len, const char *ttag, ID3v2ExtraMeta **extra_meta, int isv34)
621 {
622  AVRational time_base = {1, 1000};
623  uint32_t start, end;
624  AVChapter *chapter;
625  uint8_t *dst = NULL;
626  int taglen;
627  char tag[5];
628 
629  if (!s) {
630  /* We should probably just put the chapter data to extra_meta here
631  * and do the AVFormatContext-needing part in a separate
632  * ff_id3v2_parse_apic()-like function. */
633  av_log(NULL, AV_LOG_DEBUG, "No AVFormatContext, skipped ID3 chapter data\n");
634  return;
635  }
636 
637  if (decode_str(s, pb, 0, &dst, &len) < 0)
638  return;
639  if (len < 16)
640  return;
641 
642  start = avio_rb32(pb);
643  end = avio_rb32(pb);
644  avio_skip(pb, 8);
645 
646  chapter = avpriv_new_chapter(s, s->nb_chapters + 1, time_base, start, end, dst);
647  if (!chapter) {
648  av_free(dst);
649  return;
650  }
651 
652  len -= 16;
653  while (len > 10) {
654  if (avio_read(pb, tag, 4) < 4)
655  goto end;
656  tag[4] = 0;
657  taglen = avio_rb32(pb);
658  avio_skip(pb, 2);
659  len -= 10;
660  if (taglen < 0 || taglen > len)
661  goto end;
662  if (tag[0] == 'T')
663  read_ttag(s, pb, taglen, &chapter->metadata, tag);
664  else
665  avio_skip(pb, taglen);
666  len -= taglen;
667  }
668 
669  ff_metadata_conv(&chapter->metadata, NULL, ff_id3v2_34_metadata_conv);
670  ff_metadata_conv(&chapter->metadata, NULL, ff_id3v2_4_metadata_conv);
671 end:
672  av_free(dst);
673 }
674 
675 static void free_priv(void *obj)
676 {
677  ID3v2ExtraMetaPRIV *priv = obj;
678  av_freep(&priv->owner);
679  av_freep(&priv->data);
680  av_freep(&priv);
681 }
682 
683 static void read_priv(AVFormatContext *s, AVIOContext *pb, int taglen,
684  const char *tag, ID3v2ExtraMeta **extra_meta, int isv34)
685 {
686  ID3v2ExtraMeta *meta;
687  ID3v2ExtraMetaPRIV *priv;
688 
689  meta = av_mallocz(sizeof(*meta));
690  priv = av_mallocz(sizeof(*priv));
691 
692  if (!meta || !priv)
693  goto fail;
694 
695  if (decode_str(s, pb, ID3v2_ENCODING_ISO8859, &priv->owner, &taglen) < 0)
696  goto fail;
697 
698  priv->data = av_malloc(taglen);
699  if (!priv->data)
700  goto fail;
701 
702  priv->datasize = taglen;
703 
704  if (avio_read(pb, priv->data, priv->datasize) != priv->datasize)
705  goto fail;
706 
707  meta->tag = "PRIV";
708  meta->data = priv;
709  meta->next = *extra_meta;
710  *extra_meta = meta;
711 
712  return;
713 
714 fail:
715  if (priv)
716  free_priv(priv);
717  av_freep(&meta);
718 }
719 
720 typedef struct ID3v2EMFunc {
721  const char *tag3;
722  const char *tag4;
723  void (*read)(AVFormatContext *s, AVIOContext *pb, int taglen,
724  const char *tag, ID3v2ExtraMeta **extra_meta,
725  int isv34);
726  void (*free)(void *obj);
727 } ID3v2EMFunc;
728 
730  { "GEO", "GEOB", read_geobtag, free_geobtag },
731  { "PIC", "APIC", read_apic, free_apic },
732  { "CHAP","CHAP", read_chapter, NULL },
733  { "PRIV","PRIV", read_priv, free_priv },
734  { NULL }
735 };
736 
737 /**
738  * Get the corresponding ID3v2EMFunc struct for a tag.
739  * @param isv34 Determines if v2.2 or v2.3/4 strings are used
740  * @return A pointer to the ID3v2EMFunc struct if found, NULL otherwise.
741  */
742 static const ID3v2EMFunc *get_extra_meta_func(const char *tag, int isv34)
743 {
744  int i = 0;
745  while (id3v2_extra_meta_funcs[i].tag3) {
746  if (tag && !memcmp(tag,
747  (isv34 ? id3v2_extra_meta_funcs[i].tag4 :
748  id3v2_extra_meta_funcs[i].tag3),
749  (isv34 ? 4 : 3)))
750  return &id3v2_extra_meta_funcs[i];
751  i++;
752  }
753  return NULL;
754 }
755 
756 static void id3v2_parse(AVIOContext *pb, AVDictionary **metadata,
758  uint8_t flags, ID3v2ExtraMeta **extra_meta)
759 {
760  int isv34, unsync;
761  unsigned tlen;
762  char tag[5];
763  int64_t next, end = avio_tell(pb) + len;
764  int taghdrlen;
765  const char *reason = NULL;
766  AVIOContext pb_local;
767  AVIOContext *pbx;
768  unsigned char *buffer = NULL;
769  int buffer_size = 0;
770  const ID3v2EMFunc *extra_func = NULL;
771  unsigned char *uncompressed_buffer = NULL;
772  av_unused int uncompressed_buffer_size = 0;
773 
774  av_log(s, AV_LOG_DEBUG, "id3v2 ver:%d flags:%02X len:%d\n", version, flags, len);
775 
776  switch (version) {
777  case 2:
778  if (flags & 0x40) {
779  reason = "compression";
780  goto error;
781  }
782  isv34 = 0;
783  taghdrlen = 6;
784  break;
785 
786  case 3:
787  case 4:
788  isv34 = 1;
789  taghdrlen = 10;
790  break;
791 
792  default:
793  reason = "version";
794  goto error;
795  }
796 
797  unsync = flags & 0x80;
798 
799  if (isv34 && flags & 0x40) { /* Extended header present, just skip over it */
800  int extlen = get_size(pb, 4);
801  if (version == 4)
802  /* In v2.4 the length includes the length field we just read. */
803  extlen -= 4;
804 
805  if (extlen < 0) {
806  reason = "invalid extended header length";
807  goto error;
808  }
809  avio_skip(pb, extlen);
810  len -= extlen + 4;
811  if (len < 0) {
812  reason = "extended header too long.";
813  goto error;
814  }
815  }
816 
817  while (len >= taghdrlen) {
818  unsigned int tflags = 0;
819  int tunsync = 0;
820  int tcomp = 0;
821  int tencr = 0;
822  unsigned long av_unused dlen;
823 
824  if (isv34) {
825  if (avio_read(pb, tag, 4) < 4)
826  break;
827  tag[4] = 0;
828  if (version == 3) {
829  tlen = avio_rb32(pb);
830  } else {
831  /* some encoders incorrectly uses v3 sizes instead of syncsafe ones
832  * so check the next tag to see which one to use */
833  tlen = avio_rb32(pb);
834  if (tlen > 0x7f) {
835  if (tlen < len) {
836  int64_t cur = avio_tell(pb);
837 
838  if (ffio_ensure_seekback(pb, 2 /* tflags */ + tlen + 4 /* next tag */))
839  break;
840 
841  if (check_tag(pb, cur + 2 + size_to_syncsafe(tlen), 4) == 1)
842  tlen = size_to_syncsafe(tlen);
843  else if (check_tag(pb, cur + 2 + tlen, 4) != 1)
844  break;
845  avio_seek(pb, cur, SEEK_SET);
846  } else
847  tlen = size_to_syncsafe(tlen);
848  }
849  }
850  tflags = avio_rb16(pb);
851  tunsync = tflags & ID3v2_FLAG_UNSYNCH;
852  } else {
853  if (avio_read(pb, tag, 3) < 3)
854  break;
855  tag[3] = 0;
856  tlen = avio_rb24(pb);
857  }
858  if (tlen > (1<<28))
859  break;
860  len -= taghdrlen + tlen;
861 
862  if (len < 0)
863  break;
864 
865  next = avio_tell(pb) + tlen;
866 
867  if (!tlen) {
868  if (tag[0])
869  av_log(s, AV_LOG_DEBUG, "Invalid empty frame %s, skipping.\n",
870  tag);
871  continue;
872  }
873 
874  if (tflags & ID3v2_FLAG_DATALEN) {
875  if (tlen < 4)
876  break;
877  dlen = avio_rb32(pb);
878  tlen -= 4;
879  } else
880  dlen = tlen;
881 
882  tcomp = tflags & ID3v2_FLAG_COMPRESSION;
883  tencr = tflags & ID3v2_FLAG_ENCRYPTION;
884 
885  /* skip encrypted tags and, if no zlib, compressed tags */
886  if (tencr || (!CONFIG_ZLIB && tcomp)) {
887  const char *type;
888  if (!tcomp)
889  type = "encrypted";
890  else if (!tencr)
891  type = "compressed";
892  else
893  type = "encrypted and compressed";
894 
895  av_log(s, AV_LOG_WARNING, "Skipping %s ID3v2 frame %s.\n", type, tag);
896  avio_skip(pb, tlen);
897  /* check for text tag or supported special meta tag */
898  } else if (tag[0] == 'T' ||
899  !memcmp(tag, "USLT", 4) ||
900  (extra_meta &&
901  (extra_func = get_extra_meta_func(tag, isv34)))) {
902  pbx = pb;
903 
904  if (unsync || tunsync || tcomp) {
905  av_fast_malloc(&buffer, &buffer_size, tlen);
906  if (!buffer) {
907  av_log(s, AV_LOG_ERROR, "Failed to alloc %d bytes\n", tlen);
908  goto seek;
909  }
910  }
911  if (unsync || tunsync) {
912  int64_t end = avio_tell(pb) + tlen;
913  uint8_t *b;
914 
915  b = buffer;
916  while (avio_tell(pb) < end && b - buffer < tlen && !pb->eof_reached) {
917  *b++ = avio_r8(pb);
918  if (*(b - 1) == 0xff && avio_tell(pb) < end - 1 &&
919  b - buffer < tlen &&
920  !pb->eof_reached ) {
921  uint8_t val = avio_r8(pb);
922  *b++ = val ? val : avio_r8(pb);
923  }
924  }
925  ffio_init_context(&pb_local, buffer, b - buffer, 0, NULL, NULL, NULL,
926  NULL);
927  tlen = b - buffer;
928  pbx = &pb_local; // read from sync buffer
929  }
930 
931 #if CONFIG_ZLIB
932  if (tcomp) {
933  int err;
934 
935  av_log(s, AV_LOG_DEBUG, "Compresssed frame %s tlen=%d dlen=%ld\n", tag, tlen, dlen);
936 
937  av_fast_malloc(&uncompressed_buffer, &uncompressed_buffer_size, dlen);
938  if (!uncompressed_buffer) {
939  av_log(s, AV_LOG_ERROR, "Failed to alloc %ld bytes\n", dlen);
940  goto seek;
941  }
942 
943  if (!(unsync || tunsync)) {
944  err = avio_read(pb, buffer, tlen);
945  if (err < 0) {
946  av_log(s, AV_LOG_ERROR, "Failed to read compressed tag\n");
947  goto seek;
948  }
949  tlen = err;
950  }
951 
952  err = uncompress(uncompressed_buffer, &dlen, buffer, tlen);
953  if (err != Z_OK) {
954  av_log(s, AV_LOG_ERROR, "Failed to uncompress tag: %d\n", err);
955  goto seek;
956  }
957  ffio_init_context(&pb_local, uncompressed_buffer, dlen, 0, NULL, NULL, NULL, NULL);
958  tlen = dlen;
959  pbx = &pb_local; // read from sync buffer
960  }
961 #endif
962  if (tag[0] == 'T')
963  /* parse text tag */
964  read_ttag(s, pbx, tlen, metadata, tag);
965  else if (!memcmp(tag, "USLT", 4))
966  read_uslt(s, pbx, tlen, metadata);
967  else
968  /* parse special meta tag */
969  extra_func->read(s, pbx, tlen, tag, extra_meta, isv34);
970  } else if (!tag[0]) {
971  if (tag[1])
972  av_log(s, AV_LOG_WARNING, "invalid frame id, assuming padding\n");
973  avio_skip(pb, tlen);
974  break;
975  }
976  /* Skip to end of tag */
977 seek:
978  avio_seek(pb, next, SEEK_SET);
979  }
980 
981  /* Footer preset, always 10 bytes, skip over it */
982  if (version == 4 && flags & 0x10)
983  end += 10;
984 
985 error:
986  if (reason)
987  av_log(s, AV_LOG_INFO, "ID3v2.%d tag skipped, cannot handle %s\n",
988  version, reason);
989  avio_seek(pb, end, SEEK_SET);
990  av_free(buffer);
991  av_free(uncompressed_buffer);
992  return;
993 }
994 
995 static void id3v2_read_internal(AVIOContext *pb, AVDictionary **metadata,
996  AVFormatContext *s, const char *magic,
997  ID3v2ExtraMeta **extra_meta, int64_t max_search_size)
998 {
999  int len, ret;
1001  int found_header;
1002  int64_t start, off;
1003 
1004  if (max_search_size && max_search_size < ID3v2_HEADER_SIZE)
1005  return;
1006 
1007  start = avio_tell(pb);
1008  do {
1009  /* save the current offset in case there's nothing to read/skip */
1010  off = avio_tell(pb);
1011  if (max_search_size && off - start >= max_search_size - ID3v2_HEADER_SIZE) {
1012  avio_seek(pb, off, SEEK_SET);
1013  break;
1014  }
1015 
1016  ret = avio_read(pb, buf, ID3v2_HEADER_SIZE);
1017  if (ret != ID3v2_HEADER_SIZE) {
1018  avio_seek(pb, off, SEEK_SET);
1019  break;
1020  }
1021  found_header = ff_id3v2_match(buf, magic);
1022  if (found_header) {
1023  /* parse ID3v2 header */
1024  len = ((buf[6] & 0x7f) << 21) |
1025  ((buf[7] & 0x7f) << 14) |
1026  ((buf[8] & 0x7f) << 7) |
1027  (buf[9] & 0x7f);
1028  id3v2_parse(pb, metadata, s, len, buf[3], buf[5], extra_meta);
1029  } else {
1030  avio_seek(pb, off, SEEK_SET);
1031  }
1032  } while (found_header);
1033  ff_metadata_conv(metadata, NULL, ff_id3v2_34_metadata_conv);
1034  ff_metadata_conv(metadata, NULL, id3v2_2_metadata_conv);
1035  ff_metadata_conv(metadata, NULL, ff_id3v2_4_metadata_conv);
1036  merge_date(metadata);
1037 }
1038 
1040  const char *magic, ID3v2ExtraMeta **extra_meta)
1041 {
1042  id3v2_read_internal(pb, metadata, NULL, magic, extra_meta, 0);
1043 }
1044 
1045 void ff_id3v2_read(AVFormatContext *s, const char *magic,
1046  ID3v2ExtraMeta **extra_meta, unsigned int max_search_size)
1047 {
1048  id3v2_read_internal(s->pb, &s->metadata, s, magic, extra_meta, max_search_size);
1049 }
1050 
1052 {
1053  ID3v2ExtraMeta *current = *extra_meta, *next;
1054  const ID3v2EMFunc *extra_func;
1055 
1056  while (current) {
1057  if ((extra_func = get_extra_meta_func(current->tag, 1)))
1058  extra_func->free(current->data);
1059  next = current->next;
1060  av_freep(&current);
1061  current = next;
1062  }
1063 
1064  *extra_meta = NULL;
1065 }
1066 
1068 {
1069  ID3v2ExtraMeta *cur;
1070 
1071  for (cur = *extra_meta; cur; cur = cur->next) {
1072  ID3v2ExtraMetaAPIC *apic;
1073  AVStream *st;
1074 
1075  if (strcmp(cur->tag, "APIC"))
1076  continue;
1077  apic = cur->data;
1078 
1079  if (!(st = avformat_new_stream(s, NULL)))
1080  return AVERROR(ENOMEM);
1081 
1084  st->codec->codec_id = apic->id;
1085  av_dict_set(&st->metadata, "title", apic->description, 0);
1086  av_dict_set(&st->metadata, "comment", apic->type, 0);
1087 
1089  st->attached_pic.buf = apic->buf;
1090  st->attached_pic.data = apic->buf->data;
1092  st->attached_pic.stream_index = st->index;
1094 
1095  apic->buf = NULL;
1096  }
1097 
1098  return 0;
1099 }