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  { 0 }
59 };
60 
62  { "TCMP", "compilation" },
63  { "TDRL", "date" },
64  { "TDRC", "date" },
65  { "TDEN", "creation_time" },
66  { "TSOA", "album-sort" },
67  { "TSOP", "artist-sort" },
68  { "TSOT", "title-sort" },
69  { 0 }
70 };
71 
73  { "TAL", "album" },
74  { "TCO", "genre" },
75  { "TCP", "compilation" },
76  { "TT2", "title" },
77  { "TEN", "encoded_by" },
78  { "TP1", "artist" },
79  { "TP2", "album_artist" },
80  { "TP3", "performer" },
81  { "TRK", "track" },
82  { 0 }
83 };
84 
85 const char ff_id3v2_tags[][4] = {
86  "TALB", "TBPM", "TCOM", "TCON", "TCOP", "TDLY", "TENC", "TEXT",
87  "TFLT", "TIT1", "TIT2", "TIT3", "TKEY", "TLAN", "TLEN", "TMED",
88  "TOAL", "TOFN", "TOLY", "TOPE", "TOWN", "TPE1", "TPE2", "TPE3",
89  "TPE4", "TPOS", "TPUB", "TRCK", "TRSN", "TRSO", "TSRC", "TSSE",
90  { 0 },
91 };
92 
93 const char ff_id3v2_4_tags[][4] = {
94  "TDEN", "TDOR", "TDRC", "TDRL", "TDTG", "TIPL", "TMCL", "TMOO",
95  "TPRO", "TSOA", "TSOP", "TSOT", "TSST",
96  { 0 },
97 };
98 
99 const char ff_id3v2_3_tags[][4] = {
100  "TDAT", "TIME", "TORY", "TRDA", "TSIZ", "TYER",
101  { 0 },
102 };
103 
104 const char *ff_id3v2_picture_types[21] = {
105  "Other",
106  "32x32 pixels 'file icon'",
107  "Other file icon",
108  "Cover (front)",
109  "Cover (back)",
110  "Leaflet page",
111  "Media (e.g. label side of CD)",
112  "Lead artist/lead performer/soloist",
113  "Artist/performer",
114  "Conductor",
115  "Band/Orchestra",
116  "Composer",
117  "Lyricist/text writer",
118  "Recording Location",
119  "During recording",
120  "During performance",
121  "Movie/video screen capture",
122  "A bright coloured fish",
123  "Illustration",
124  "Band/artist logotype",
125  "Publisher/Studio logotype",
126 };
127 
129  { "image/gif", AV_CODEC_ID_GIF },
130  { "image/jpeg", AV_CODEC_ID_MJPEG },
131  { "image/jpg", AV_CODEC_ID_MJPEG },
132  { "image/png", AV_CODEC_ID_PNG },
133  { "image/tiff", AV_CODEC_ID_TIFF },
134  { "image/bmp", AV_CODEC_ID_BMP },
135  { "JPG", AV_CODEC_ID_MJPEG }, /* ID3v2.2 */
136  { "PNG", AV_CODEC_ID_PNG }, /* ID3v2.2 */
137  { "", AV_CODEC_ID_NONE },
138 };
139 
140 int ff_id3v2_match(const uint8_t *buf, const char *magic)
141 {
142  return buf[0] == magic[0] &&
143  buf[1] == magic[1] &&
144  buf[2] == magic[2] &&
145  buf[3] != 0xff &&
146  buf[4] != 0xff &&
147  (buf[6] & 0x80) == 0 &&
148  (buf[7] & 0x80) == 0 &&
149  (buf[8] & 0x80) == 0 &&
150  (buf[9] & 0x80) == 0;
151 }
152 
154 {
155  int len = ((buf[6] & 0x7f) << 21) +
156  ((buf[7] & 0x7f) << 14) +
157  ((buf[8] & 0x7f) << 7) +
158  (buf[9] & 0x7f) +
160  if (buf[5] & 0x10)
161  len += ID3v2_HEADER_SIZE;
162  return len;
163 }
164 
165 static unsigned int get_size(AVIOContext *s, int len)
166 {
167  int v = 0;
168  while (len--)
169  v = (v << 7) + (avio_r8(s) & 0x7F);
170  return v;
171 }
172 
173 /**
174  * Free GEOB type extra metadata.
175  */
176 static void free_geobtag(void *obj)
177 {
178  ID3v2ExtraMetaGEOB *geob = obj;
179  av_free(geob->mime_type);
180  av_free(geob->file_name);
181  av_free(geob->description);
182  av_free(geob->data);
183  av_free(geob);
184 }
185 
186 /**
187  * Decode characters to UTF-8 according to encoding type. The decoded buffer is
188  * always null terminated. Stop reading when either *maxread bytes are read from
189  * pb or U+0000 character is found.
190  *
191  * @param dst Pointer where the address of the buffer with the decoded bytes is
192  * stored. Buffer must be freed by caller.
193  * @param maxread Pointer to maximum number of characters to read from the
194  * AVIOContext. After execution the value is decremented by the number of bytes
195  * actually read.
196  * @returns 0 if no error occurred, dst is uninitialized on error
197  */
198 static int decode_str(AVFormatContext *s, AVIOContext *pb, int encoding,
199  uint8_t **dst, int *maxread)
200 {
201  int ret;
202  uint8_t tmp;
203  uint32_t ch = 1;
204  int left = *maxread;
205  unsigned int (*get)(AVIOContext*) = avio_rb16;
206  AVIOContext *dynbuf;
207 
208  if ((ret = avio_open_dyn_buf(&dynbuf)) < 0) {
209  av_log(s, AV_LOG_ERROR, "Error opening memory stream\n");
210  return ret;
211  }
212 
213  switch (encoding) {
215  while (left && ch) {
216  ch = avio_r8(pb);
217  PUT_UTF8(ch, tmp, avio_w8(dynbuf, tmp);)
218  left--;
219  }
220  break;
221 
223  if ((left -= 2) < 0) {
224  av_log(s, AV_LOG_ERROR, "Cannot read BOM value, input too short\n");
225  avio_close_dyn_buf(dynbuf, dst);
226  av_freep(dst);
227  return AVERROR_INVALIDDATA;
228  }
229  switch (avio_rb16(pb)) {
230  case 0xfffe:
231  get = avio_rl16;
232  case 0xfeff:
233  break;
234  default:
235  av_log(s, AV_LOG_ERROR, "Incorrect BOM value\n");
236  avio_close_dyn_buf(dynbuf, dst);
237  av_freep(dst);
238  *maxread = left;
239  return AVERROR_INVALIDDATA;
240  }
241  // fall-through
242 
244  while ((left > 1) && ch) {
245  GET_UTF16(ch, ((left -= 2) >= 0 ? get(pb) : 0), break;)
246  PUT_UTF8(ch, tmp, avio_w8(dynbuf, tmp);)
247  }
248  if (left < 0)
249  left += 2; /* did not read last char from pb */
250  break;
251 
252  case ID3v2_ENCODING_UTF8:
253  while (left && ch) {
254  ch = avio_r8(pb);
255  avio_w8(dynbuf, ch);
256  left--;
257  }
258  break;
259  default:
260  av_log(s, AV_LOG_WARNING, "Unknown encoding\n");
261  }
262 
263  if (ch)
264  avio_w8(dynbuf, 0);
265 
266  avio_close_dyn_buf(dynbuf, dst);
267  *maxread = left;
268 
269  return 0;
270 }
271 
272 /**
273  * Parse a text tag.
274  */
275 static void read_ttag(AVFormatContext *s, AVIOContext *pb, int taglen,
276  AVDictionary **metadata, const char *key)
277 {
278  uint8_t *dst;
279  int encoding, dict_flags = AV_DICT_DONT_OVERWRITE | AV_DICT_DONT_STRDUP_VAL;
280  unsigned genre;
281 
282  if (taglen < 1)
283  return;
284 
285  encoding = avio_r8(pb);
286  taglen--; /* account for encoding type byte */
287 
288  if (decode_str(s, pb, encoding, &dst, &taglen) < 0) {
289  av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", key);
290  return;
291  }
292 
293  if (!(strcmp(key, "TCON") && strcmp(key, "TCO")) &&
294  (sscanf(dst, "(%d)", &genre) == 1 || sscanf(dst, "%d", &genre) == 1) &&
295  genre <= ID3v1_GENRE_MAX) {
296  av_freep(&dst);
297  dst = av_strdup(ff_id3v1_genre_str[genre]);
298  } else if (!(strcmp(key, "TXXX") && strcmp(key, "TXX"))) {
299  /* dst now contains the key, need to get value */
300  key = dst;
301  if (decode_str(s, pb, encoding, &dst, &taglen) < 0) {
302  av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", key);
303  av_freep(&key);
304  return;
305  }
306  dict_flags |= AV_DICT_DONT_STRDUP_KEY;
307  } else if (!*dst)
308  av_freep(&dst);
309 
310  if (dst)
311  av_dict_set(metadata, key, dst, dict_flags);
312 }
313 
314 /**
315  * Parse GEOB tag into a ID3v2ExtraMetaGEOB struct.
316  */
317 static void read_geobtag(AVFormatContext *s, AVIOContext *pb, int taglen,
318  char *tag, ID3v2ExtraMeta **extra_meta, int isv34)
319 {
320  ID3v2ExtraMetaGEOB *geob_data = NULL;
321  ID3v2ExtraMeta *new_extra = NULL;
322  char encoding;
323  unsigned int len;
324 
325  if (taglen < 1)
326  return;
327 
328  geob_data = av_mallocz(sizeof(ID3v2ExtraMetaGEOB));
329  if (!geob_data) {
330  av_log(s, AV_LOG_ERROR, "Failed to alloc %"SIZE_SPECIFIER" bytes\n",
331  sizeof(ID3v2ExtraMetaGEOB));
332  return;
333  }
334 
335  new_extra = av_mallocz(sizeof(ID3v2ExtraMeta));
336  if (!new_extra) {
337  av_log(s, AV_LOG_ERROR, "Failed to alloc %"SIZE_SPECIFIER" bytes\n",
338  sizeof(ID3v2ExtraMeta));
339  goto fail;
340  }
341 
342  /* read encoding type byte */
343  encoding = avio_r8(pb);
344  taglen--;
345 
346  /* read MIME type (always ISO-8859) */
347  if (decode_str(s, pb, ID3v2_ENCODING_ISO8859, &geob_data->mime_type,
348  &taglen) < 0 ||
349  taglen <= 0)
350  goto fail;
351 
352  /* read file name */
353  if (decode_str(s, pb, encoding, &geob_data->file_name, &taglen) < 0 ||
354  taglen <= 0)
355  goto fail;
356 
357  /* read content description */
358  if (decode_str(s, pb, encoding, &geob_data->description, &taglen) < 0 ||
359  taglen < 0)
360  goto fail;
361 
362  if (taglen) {
363  /* save encapsulated binary data */
364  geob_data->data = av_malloc(taglen);
365  if (!geob_data->data) {
366  av_log(s, AV_LOG_ERROR, "Failed to alloc %d bytes\n", taglen);
367  goto fail;
368  }
369  if ((len = avio_read(pb, geob_data->data, taglen)) < taglen)
371  "Error reading GEOB frame, data truncated.\n");
372  geob_data->datasize = len;
373  } else {
374  geob_data->data = NULL;
375  geob_data->datasize = 0;
376  }
377 
378  /* add data to the list */
379  new_extra->tag = "GEOB";
380  new_extra->data = geob_data;
381  new_extra->next = *extra_meta;
382  *extra_meta = new_extra;
383 
384  return;
385 
386 fail:
387  av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", tag);
388  free_geobtag(geob_data);
389  av_free(new_extra);
390  return;
391 }
392 
393 static int is_number(const char *str)
394 {
395  while (*str >= '0' && *str <= '9')
396  str++;
397  return !*str;
398 }
399 
401 {
403  if ((t = av_dict_get(m, tag, NULL, AV_DICT_MATCH_CASE)) &&
404  strlen(t->value) == 4 && is_number(t->value))
405  return t;
406  return NULL;
407 }
408 
409 static void merge_date(AVDictionary **m)
410 {
412  char date[17] = { 0 }; // YYYY-MM-DD hh:mm
413 
414  if (!(t = get_date_tag(*m, "TYER")) &&
415  !(t = get_date_tag(*m, "TYE")))
416  return;
417  av_strlcpy(date, t->value, 5);
418  av_dict_set(m, "TYER", NULL, 0);
419  av_dict_set(m, "TYE", NULL, 0);
420 
421  if (!(t = get_date_tag(*m, "TDAT")) &&
422  !(t = get_date_tag(*m, "TDA")))
423  goto finish;
424  snprintf(date + 4, sizeof(date) - 4, "-%.2s-%.2s", t->value + 2, t->value);
425  av_dict_set(m, "TDAT", NULL, 0);
426  av_dict_set(m, "TDA", NULL, 0);
427 
428  if (!(t = get_date_tag(*m, "TIME")) &&
429  !(t = get_date_tag(*m, "TIM")))
430  goto finish;
431  snprintf(date + 10, sizeof(date) - 10,
432  " %.2s:%.2s", t->value, t->value + 2);
433  av_dict_set(m, "TIME", NULL, 0);
434  av_dict_set(m, "TIM", NULL, 0);
435 
436 finish:
437  if (date[0])
438  av_dict_set(m, "date", date, 0);
439 }
440 
441 static void free_apic(void *obj)
442 {
443  ID3v2ExtraMetaAPIC *apic = obj;
444  av_buffer_unref(&apic->buf);
445  av_freep(&apic->description);
446  av_freep(&apic);
447 }
448 
449 static void read_apic(AVFormatContext *s, AVIOContext *pb, int taglen,
450  char *tag, ID3v2ExtraMeta **extra_meta, int isv34)
451 {
452  int enc, pic_type;
453  char mimetype[64];
454  const CodecMime *mime = ff_id3v2_mime_tags;
455  enum AVCodecID id = AV_CODEC_ID_NONE;
456  ID3v2ExtraMetaAPIC *apic = NULL;
457  ID3v2ExtraMeta *new_extra = NULL;
458  int64_t end = avio_tell(pb) + taglen;
459 
460  if (taglen <= 4)
461  goto fail;
462 
463  new_extra = av_mallocz(sizeof(*new_extra));
464  apic = av_mallocz(sizeof(*apic));
465  if (!new_extra || !apic)
466  goto fail;
467 
468  enc = avio_r8(pb);
469  taglen--;
470 
471  /* mimetype */
472  if (isv34) {
473  taglen -= avio_get_str(pb, taglen, mimetype, sizeof(mimetype));
474  } else {
475  avio_read(pb, mimetype, 3);
476  mimetype[3] = 0;
477  }
478  while (mime->id != AV_CODEC_ID_NONE) {
479  if (!av_strncasecmp(mime->str, mimetype, sizeof(mimetype))) {
480  id = mime->id;
481  break;
482  }
483  mime++;
484  }
485  if (id == AV_CODEC_ID_NONE) {
487  "Unknown attached picture mimetype: %s, skipping.\n", mimetype);
488  goto fail;
489  }
490  apic->id = id;
491 
492  /* picture type */
493  pic_type = avio_r8(pb);
494  taglen--;
495  if (pic_type < 0 || pic_type >= FF_ARRAY_ELEMS(ff_id3v2_picture_types)) {
496  av_log(s, AV_LOG_WARNING, "Unknown attached picture type %d.\n",
497  pic_type);
498  pic_type = 0;
499  }
500  apic->type = ff_id3v2_picture_types[pic_type];
501 
502  /* description and picture data */
503  if (decode_str(s, pb, enc, &apic->description, &taglen) < 0) {
504  av_log(s, AV_LOG_ERROR,
505  "Error decoding attached picture description.\n");
506  goto fail;
507  }
508 
510  if (!apic->buf || !taglen || avio_read(pb, apic->buf->data, taglen) != taglen)
511  goto fail;
512  memset(apic->buf->data + taglen, 0, FF_INPUT_BUFFER_PADDING_SIZE);
513 
514  new_extra->tag = "APIC";
515  new_extra->data = apic;
516  new_extra->next = *extra_meta;
517  *extra_meta = new_extra;
518 
519  return;
520 
521 fail:
522  if (apic)
523  free_apic(apic);
524  av_freep(&new_extra);
525  avio_seek(pb, end, SEEK_SET);
526 }
527 
528 static void read_chapter(AVFormatContext *s, AVIOContext *pb, int len, char *ttag, ID3v2ExtraMeta **extra_meta, int isv34)
529 {
530  AVRational time_base = {1, 1000};
531  uint32_t start, end;
532  AVChapter *chapter;
533  uint8_t *dst = NULL;
534  int taglen;
535  char tag[5];
536 
537  if (!s) {
538  /* We should probably just put the chapter data to extra_meta here
539  * and do the AVFormatContext-needing part in a separate
540  * ff_id3v2_parse_apic()-like function. */
541  av_log(NULL, AV_LOG_DEBUG, "No AVFormatContext, skipped ID3 chapter data\n");
542  return;
543  }
544 
545  if (decode_str(s, pb, 0, &dst, &len) < 0)
546  return;
547  if (len < 16)
548  return;
549 
550  start = avio_rb32(pb);
551  end = avio_rb32(pb);
552  avio_skip(pb, 8);
553 
554  chapter = avpriv_new_chapter(s, s->nb_chapters + 1, time_base, start, end, dst);
555  if (!chapter) {
556  av_free(dst);
557  return;
558  }
559 
560  len -= 16;
561  while (len > 10) {
562  if (avio_read(pb, tag, 4) < 4)
563  goto end;
564  tag[4] = 0;
565  taglen = avio_rb32(pb);
566  avio_skip(pb, 2);
567  len -= 10;
568  if (taglen < 0 || taglen > len)
569  goto end;
570  if (tag[0] == 'T')
571  read_ttag(s, pb, taglen, &chapter->metadata, tag);
572  else
573  avio_skip(pb, taglen);
574  len -= taglen;
575  }
576 
577  ff_metadata_conv(&chapter->metadata, NULL, ff_id3v2_34_metadata_conv);
578  ff_metadata_conv(&chapter->metadata, NULL, ff_id3v2_4_metadata_conv);
579 end:
580  av_free(dst);
581 }
582 
583 static void free_priv(void *obj)
584 {
585  ID3v2ExtraMetaPRIV *priv = obj;
586  av_freep(&priv->owner);
587  av_freep(&priv->data);
588  av_freep(&priv);
589 }
590 
591 static void read_priv(AVFormatContext *s, AVIOContext *pb, int taglen,
592  char *tag, ID3v2ExtraMeta **extra_meta, int isv34)
593 {
594  ID3v2ExtraMeta *meta;
595  ID3v2ExtraMetaPRIV *priv;
596 
597  meta = av_mallocz(sizeof(*meta));
598  priv = av_mallocz(sizeof(*priv));
599 
600  if (!meta || !priv)
601  goto fail;
602 
603  if (decode_str(s, pb, ID3v2_ENCODING_ISO8859, &priv->owner, &taglen) < 0)
604  goto fail;
605 
606  priv->data = av_malloc(taglen);
607  if (!priv->data)
608  goto fail;
609 
610  priv->datasize = taglen;
611 
612  if (avio_read(pb, priv->data, priv->datasize) != priv->datasize)
613  goto fail;
614 
615  meta->tag = "PRIV";
616  meta->data = priv;
617  meta->next = *extra_meta;
618  *extra_meta = meta;
619 
620  return;
621 
622 fail:
623  if (priv)
624  free_priv(priv);
625  av_freep(&meta);
626 }
627 
628 typedef struct ID3v2EMFunc {
629  const char *tag3;
630  const char *tag4;
631  void (*read)(AVFormatContext *, AVIOContext *, int, char *,
632  ID3v2ExtraMeta **, int isv34);
633  void (*free)(void *obj);
634 } ID3v2EMFunc;
635 
637  { "GEO", "GEOB", read_geobtag, free_geobtag },
638  { "PIC", "APIC", read_apic, free_apic },
639  { "CHAP","CHAP", read_chapter, NULL },
640  { "PRIV","PRIV", read_priv, free_priv },
641  { NULL }
642 };
643 
644 /**
645  * Get the corresponding ID3v2EMFunc struct for a tag.
646  * @param isv34 Determines if v2.2 or v2.3/4 strings are used
647  * @return A pointer to the ID3v2EMFunc struct if found, NULL otherwise.
648  */
649 static const ID3v2EMFunc *get_extra_meta_func(const char *tag, int isv34)
650 {
651  int i = 0;
652  while (id3v2_extra_meta_funcs[i].tag3) {
653  if (tag && !memcmp(tag,
654  (isv34 ? id3v2_extra_meta_funcs[i].tag4 :
655  id3v2_extra_meta_funcs[i].tag3),
656  (isv34 ? 4 : 3)))
657  return &id3v2_extra_meta_funcs[i];
658  i++;
659  }
660  return NULL;
661 }
662 
663 static void id3v2_parse(AVIOContext *pb, AVDictionary **metadata,
665  uint8_t flags, ID3v2ExtraMeta **extra_meta)
666 {
667  int isv34, unsync;
668  unsigned tlen;
669  char tag[5];
670  int64_t next, end = avio_tell(pb) + len;
671  int taghdrlen;
672  const char *reason = NULL;
673  AVIOContext pb_local;
674  AVIOContext *pbx;
675  unsigned char *buffer = NULL;
676  int buffer_size = 0;
677  const ID3v2EMFunc *extra_func = NULL;
678  unsigned char *uncompressed_buffer = NULL;
679  int uncompressed_buffer_size = 0;
680 
681  av_log(s, AV_LOG_DEBUG, "id3v2 ver:%d flags:%02X len:%d\n", version, flags, len);
682 
683  switch (version) {
684  case 2:
685  if (flags & 0x40) {
686  reason = "compression";
687  goto error;
688  }
689  isv34 = 0;
690  taghdrlen = 6;
691  break;
692 
693  case 3:
694  case 4:
695  isv34 = 1;
696  taghdrlen = 10;
697  break;
698 
699  default:
700  reason = "version";
701  goto error;
702  }
703 
704  unsync = flags & 0x80;
705 
706  if (isv34 && flags & 0x40) { /* Extended header present, just skip over it */
707  int extlen = get_size(pb, 4);
708  if (version == 4)
709  /* In v2.4 the length includes the length field we just read. */
710  extlen -= 4;
711 
712  if (extlen < 0) {
713  reason = "invalid extended header length";
714  goto error;
715  }
716  avio_skip(pb, extlen);
717  len -= extlen + 4;
718  if (len < 0) {
719  reason = "extended header too long.";
720  goto error;
721  }
722  }
723 
724  while (len >= taghdrlen) {
725  unsigned int tflags = 0;
726  int tunsync = 0;
727  int tcomp = 0;
728  int tencr = 0;
729  unsigned long dlen;
730 
731  if (isv34) {
732  if (avio_read(pb, tag, 4) < 4)
733  break;
734  tag[4] = 0;
735  if (version == 3) {
736  tlen = avio_rb32(pb);
737  } else
738  tlen = get_size(pb, 4);
739  tflags = avio_rb16(pb);
740  tunsync = tflags & ID3v2_FLAG_UNSYNCH;
741  } else {
742  if (avio_read(pb, tag, 3) < 3)
743  break;
744  tag[3] = 0;
745  tlen = avio_rb24(pb);
746  }
747  if (tlen > (1<<28))
748  break;
749  len -= taghdrlen + tlen;
750 
751  if (len < 0)
752  break;
753 
754  next = avio_tell(pb) + tlen;
755 
756  if (!tlen) {
757  if (tag[0])
758  av_log(s, AV_LOG_DEBUG, "Invalid empty frame %s, skipping.\n",
759  tag);
760  continue;
761  }
762 
763  if (tflags & ID3v2_FLAG_DATALEN) {
764  if (tlen < 4)
765  break;
766  dlen = avio_rb32(pb);
767  tlen -= 4;
768  } else
769  dlen = tlen;
770 
771  tcomp = tflags & ID3v2_FLAG_COMPRESSION;
772  tencr = tflags & ID3v2_FLAG_ENCRYPTION;
773 
774  /* skip encrypted tags and, if no zlib, compressed tags */
775  if (tencr || (!CONFIG_ZLIB && tcomp)) {
776  const char *type;
777  if (!tcomp)
778  type = "encrypted";
779  else if (!tencr)
780  type = "compressed";
781  else
782  type = "encrypted and compressed";
783 
784  av_log(s, AV_LOG_WARNING, "Skipping %s ID3v2 frame %s.\n", type, tag);
785  avio_skip(pb, tlen);
786  /* check for text tag or supported special meta tag */
787  } else if (tag[0] == 'T' ||
788  (extra_meta &&
789  (extra_func = get_extra_meta_func(tag, isv34)))) {
790  pbx = pb;
791 
792  if (unsync || tunsync || tcomp) {
793  av_fast_malloc(&buffer, &buffer_size, tlen);
794  if (!buffer) {
795  av_log(s, AV_LOG_ERROR, "Failed to alloc %d bytes\n", tlen);
796  goto seek;
797  }
798  }
799  if (unsync || tunsync) {
800  int64_t end = avio_tell(pb) + tlen;
801  uint8_t *b;
802 
803  b = buffer;
804  while (avio_tell(pb) < end && b - buffer < tlen && !pb->eof_reached) {
805  *b++ = avio_r8(pb);
806  if (*(b - 1) == 0xff && avio_tell(pb) < end - 1 &&
807  b - buffer < tlen &&
808  !pb->eof_reached ) {
809  uint8_t val = avio_r8(pb);
810  *b++ = val ? val : avio_r8(pb);
811  }
812  }
813  ffio_init_context(&pb_local, buffer, b - buffer, 0, NULL, NULL, NULL,
814  NULL);
815  tlen = b - buffer;
816  pbx = &pb_local; // read from sync buffer
817  }
818 
819 #if CONFIG_ZLIB
820  if (tcomp) {
821  int err;
822 
823  av_log(s, AV_LOG_DEBUG, "Compresssed frame %s tlen=%d dlen=%ld\n", tag, tlen, dlen);
824 
825  av_fast_malloc(&uncompressed_buffer, &uncompressed_buffer_size, dlen);
826  if (!uncompressed_buffer) {
827  av_log(s, AV_LOG_ERROR, "Failed to alloc %ld bytes\n", dlen);
828  goto seek;
829  }
830 
831  if (!(unsync || tunsync)) {
832  err = avio_read(pb, buffer, tlen);
833  if (err < 0) {
834  av_log(s, AV_LOG_ERROR, "Failed to read compressed tag\n");
835  goto seek;
836  }
837  tlen = err;
838  }
839 
840  err = uncompress(uncompressed_buffer, &dlen, buffer, tlen);
841  if (err != Z_OK) {
842  av_log(s, AV_LOG_ERROR, "Failed to uncompress tag: %d\n", err);
843  goto seek;
844  }
845  ffio_init_context(&pb_local, uncompressed_buffer, dlen, 0, NULL, NULL, NULL, NULL);
846  tlen = dlen;
847  pbx = &pb_local; // read from sync buffer
848  }
849 #endif
850  if (tag[0] == 'T')
851  /* parse text tag */
852  read_ttag(s, pbx, tlen, metadata, tag);
853  else
854  /* parse special meta tag */
855  extra_func->read(s, pbx, tlen, tag, extra_meta, isv34);
856  } else if (!tag[0]) {
857  if (tag[1])
858  av_log(s, AV_LOG_WARNING, "invalid frame id, assuming padding\n");
859  avio_skip(pb, tlen);
860  break;
861  }
862  /* Skip to end of tag */
863 seek:
864  avio_seek(pb, next, SEEK_SET);
865  }
866 
867  /* Footer preset, always 10 bytes, skip over it */
868  if (version == 4 && flags & 0x10)
869  end += 10;
870 
871 error:
872  if (reason)
873  av_log(s, AV_LOG_INFO, "ID3v2.%d tag skipped, cannot handle %s\n",
874  version, reason);
875  avio_seek(pb, end, SEEK_SET);
876  av_free(buffer);
877  av_free(uncompressed_buffer);
878  return;
879 }
880 
881 static void id3v2_read_internal(AVIOContext *pb, AVDictionary **metadata,
882  AVFormatContext *s, const char *magic,
883  ID3v2ExtraMeta **extra_meta, int64_t max_search_size)
884 {
885  int len, ret;
887  int found_header;
888  int64_t start, off;
889 
890  if (max_search_size && max_search_size < ID3v2_HEADER_SIZE)
891  return;
892 
893  start = avio_tell(pb);
894  do {
895  /* save the current offset in case there's nothing to read/skip */
896  off = avio_tell(pb);
897  if (max_search_size && off - start >= max_search_size - ID3v2_HEADER_SIZE) {
898  avio_seek(pb, off, SEEK_SET);
899  break;
900  }
901 
902  ret = avio_read(pb, buf, ID3v2_HEADER_SIZE);
903  if (ret != ID3v2_HEADER_SIZE) {
904  avio_seek(pb, off, SEEK_SET);
905  break;
906  }
907  found_header = ff_id3v2_match(buf, magic);
908  if (found_header) {
909  /* parse ID3v2 header */
910  len = ((buf[6] & 0x7f) << 21) |
911  ((buf[7] & 0x7f) << 14) |
912  ((buf[8] & 0x7f) << 7) |
913  (buf[9] & 0x7f);
914  id3v2_parse(pb, metadata, s, len, buf[3], buf[5], extra_meta);
915  } else {
916  avio_seek(pb, off, SEEK_SET);
917  }
918  } while (found_header);
919  ff_metadata_conv(metadata, NULL, ff_id3v2_34_metadata_conv);
920  ff_metadata_conv(metadata, NULL, id3v2_2_metadata_conv);
921  ff_metadata_conv(metadata, NULL, ff_id3v2_4_metadata_conv);
922  merge_date(metadata);
923 }
924 
926  const char *magic, ID3v2ExtraMeta **extra_meta)
927 {
928  id3v2_read_internal(pb, metadata, NULL, magic, extra_meta, 0);
929 }
930 
931 void ff_id3v2_read(AVFormatContext *s, const char *magic,
932  ID3v2ExtraMeta **extra_meta, unsigned int max_search_size)
933 {
934  id3v2_read_internal(s->pb, &s->metadata, s, magic, extra_meta, max_search_size);
935 }
936 
938 {
939  ID3v2ExtraMeta *current = *extra_meta, *next;
940  const ID3v2EMFunc *extra_func;
941 
942  while (current) {
943  if ((extra_func = get_extra_meta_func(current->tag, 1)))
944  extra_func->free(current->data);
945  next = current->next;
946  av_freep(&current);
947  current = next;
948  }
949 
950  *extra_meta = NULL;
951 }
952 
954 {
955  ID3v2ExtraMeta *cur;
956 
957  for (cur = *extra_meta; cur; cur = cur->next) {
958  ID3v2ExtraMetaAPIC *apic;
959  AVStream *st;
960 
961  if (strcmp(cur->tag, "APIC"))
962  continue;
963  apic = cur->data;
964 
965  if (!(st = avformat_new_stream(s, NULL)))
966  return AVERROR(ENOMEM);
967 
970  st->codec->codec_id = apic->id;
971  av_dict_set(&st->metadata, "title", apic->description, 0);
972  av_dict_set(&st->metadata, "comment", apic->type, 0);
973 
975  st->attached_pic.buf = apic->buf;
976  st->attached_pic.data = apic->buf->data;
978  st->attached_pic.stream_index = st->index;
980 
981  apic->buf = NULL;
982  }
983 
984  return 0;
985 }