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
45 const 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
64 const 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
75 static 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
88 const 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
96 const 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
102 const char ff_id3v2_3_tags[][4] = {
103 "TDAT", "TIME", "TORY", "TRDA", "TSIZ", "TYER",
104 { 0 },
105 };
106
107 const 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
131 const 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
ff_id3v2_match(const uint8_t * buf,const char * magic)143 int 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
ff_id3v2_tag_len(const uint8_t * buf)156 int 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
get_size(AVIOContext * s,int len)168 static 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
size_to_syncsafe(unsigned int size)176 static 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 */
is_tag(const char * buf,unsigned int len)186 static 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 */
check_tag(AVIOContext * s,int offset,unsigned int len)204 static 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 */
free_geobtag(void * obj)221 static 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 }
229
230 /**
231 * Decode characters to UTF-8 according to encoding type. The decoded buffer is
232 * always null terminated. Stop reading when either *maxread bytes are read from
233 * pb or U+0000 character is found.
234 *
235 * @param dst Pointer where the address of the buffer with the decoded bytes is
236 * stored. Buffer must be freed by caller.
237 * @param maxread Pointer to maximum number of characters to read from the
238 * AVIOContext. After execution the value is decremented by the number of bytes
239 * actually read.
240 * @returns 0 if no error occurred, dst is uninitialized on error
241 */
decode_str(AVFormatContext * s,AVIOContext * pb,int encoding,uint8_t ** dst,int * maxread)242 static int decode_str(AVFormatContext *s, AVIOContext *pb, int encoding,
243 uint8_t **dst, int *maxread)
244 {
245 int ret;
246 uint8_t tmp;
247 uint32_t ch = 1;
248 int left = *maxread;
249 unsigned int (*get)(AVIOContext*) = avio_rb16;
250 AVIOContext *dynbuf;
251
252 if ((ret = avio_open_dyn_buf(&dynbuf)) < 0) {
253 av_log(s, AV_LOG_ERROR, "Error opening memory stream\n");
254 return ret;
255 }
256
257 switch (encoding) {
258 case ID3v2_ENCODING_ISO8859:
259 while (left && ch) {
260 ch = avio_r8(pb);
261 PUT_UTF8(ch, tmp, avio_w8(dynbuf, tmp);)
262 left--;
263 }
264 break;
265
266 case ID3v2_ENCODING_UTF16BOM:
267 if ((left -= 2) < 0) {
268 av_log(s, AV_LOG_ERROR, "Cannot read BOM value, input too short\n");
269 ffio_free_dyn_buf(&dynbuf);
270 *dst = NULL;
271 return AVERROR_INVALIDDATA;
272 }
273 switch (avio_rb16(pb)) {
274 case 0xfffe:
275 get = avio_rl16;
276 case 0xfeff:
277 break;
278 default:
279 av_log(s, AV_LOG_ERROR, "Incorrect BOM value\n");
280 ffio_free_dyn_buf(&dynbuf);
281 *dst = NULL;
282 *maxread = left;
283 return AVERROR_INVALIDDATA;
284 }
285 // fall-through
286
287 case ID3v2_ENCODING_UTF16BE:
288 while ((left > 1) && ch) {
289 GET_UTF16(ch, ((left -= 2) >= 0 ? get(pb) : 0), break;)
290 PUT_UTF8(ch, tmp, avio_w8(dynbuf, tmp);)
291 }
292 if (left < 0)
293 left += 2; /* did not read last char from pb */
294 break;
295
296 case ID3v2_ENCODING_UTF8:
297 while (left && ch) {
298 ch = avio_r8(pb);
299 avio_w8(dynbuf, ch);
300 left--;
301 }
302 break;
303 default:
304 av_log(s, AV_LOG_WARNING, "Unknown encoding\n");
305 }
306
307 if (ch)
308 avio_w8(dynbuf, 0);
309
310 avio_close_dyn_buf(dynbuf, dst);
311 *maxread = left;
312
313 return 0;
314 }
315
316 #ifdef OHOS_OPT_COMPAT
317 #include <iconv.h>
iso8859_convert_utf8(char * input,size_t inputlen,char * output,size_t outputlen)318 static void iso8859_convert_utf8 (char *input, size_t inputlen, char *output, size_t outputlen)
319 {
320 size_t inbuferlen = inputlen;
321 size_t outbuferlen = outputlen;
322 iconv_t cd = iconv_open("ISO-8859-1", "UTF-8");
323 if (cd != (iconv_t)-1) {
324 size_t ret = iconv(cd, &input, (size_t *)&inbuferlen, &output, (size_t *)&outbuferlen);
325 (void)ret;
326 iconv_close(cd);
327 return;
328 }
329 }
330 #endif
331
332 /**
333 * Parse a text tag.
334 */
read_ttag(AVFormatContext * s,AVIOContext * pb,int taglen,AVDictionary ** metadata,const char * key)335 static void read_ttag(AVFormatContext *s, AVIOContext *pb, int taglen,
336 AVDictionary **metadata, const char *key)
337 {
338 uint8_t *dst;
339 int encoding, dict_flags = AV_DICT_DONT_OVERWRITE | AV_DICT_DONT_STRDUP_VAL;
340 unsigned genre;
341
342 if (taglen < 1)
343 return;
344
345 encoding = avio_r8(pb);
346 taglen--; /* account for encoding type byte */
347
348 if (decode_str(s, pb, encoding, &dst, &taglen) < 0) {
349 av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", key);
350 return;
351 }
352
353 if (!(strcmp(key, "TCON") && strcmp(key, "TCO")) &&
354 (sscanf(dst, "(%d)", &genre) == 1 || sscanf(dst, "%d", &genre) == 1) &&
355 genre <= ID3v1_GENRE_MAX) {
356 av_freep(&dst);
357 dst = av_strdup(ff_id3v1_genre_str[genre]);
358 } else if (!(strcmp(key, "TXXX") && strcmp(key, "TXX"))) {
359 /* dst now contains the key, need to get value */
360 key = dst;
361 if (decode_str(s, pb, encoding, &dst, &taglen) < 0) {
362 av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", key);
363 av_freep(&key);
364 return;
365 }
366 dict_flags |= AV_DICT_DONT_STRDUP_KEY;
367 } else if (!*dst)
368 av_freep(&dst);
369
370 #ifdef OHOS_OPT_COMPAT
371 if (dst) {
372 if (encoding == ID3v2_ENCODING_ISO8859) {
373 const int utf8len = 256;
374 char *utf8 = av_malloc(utf8len + 1);
375 utf8[utf8len] = '\0';
376 iso8859_convert_utf8(dst, strlen(dst), utf8, utf8len);
377 av_dict_set(metadata, key, utf8, dict_flags);
378 av_freep(&dst);
379 } else {
380 av_dict_set(metadata, key, dst, dict_flags);
381 }
382 }
383 #else
384 if (dst)
385 av_dict_set(metadata, key, dst, dict_flags);
386 #endif
387 }
388
read_uslt(AVFormatContext * s,AVIOContext * pb,int taglen,AVDictionary ** metadata)389 static void read_uslt(AVFormatContext *s, AVIOContext *pb, int taglen,
390 AVDictionary **metadata)
391 {
392 uint8_t lang[4];
393 uint8_t *descriptor = NULL; // 'Content descriptor'
394 uint8_t *text;
395 char *key;
396 int encoding;
397 int ok = 0;
398
399 if (taglen < 1)
400 goto error;
401
402 encoding = avio_r8(pb);
403 taglen--;
404
405 if (avio_read(pb, lang, 3) < 3)
406 goto error;
407 lang[3] = '\0';
408 taglen -= 3;
409
410 if (decode_str(s, pb, encoding, &descriptor, &taglen) < 0)
411 goto error;
412
413 if (decode_str(s, pb, encoding, &text, &taglen) < 0)
414 goto error;
415
416 // FFmpeg does not support hierarchical metadata, so concatenate the keys.
417 key = av_asprintf("lyrics-%s%s%s", descriptor[0] ? (char *)descriptor : "",
418 descriptor[0] ? "-" : "",
419 lang);
420 if (!key) {
421 av_free(text);
422 goto error;
423 }
424
425 av_dict_set(metadata, key, text,
426 AV_DICT_DONT_STRDUP_KEY | AV_DICT_DONT_STRDUP_VAL);
427
428 ok = 1;
429 error:
430 if (!ok)
431 av_log(s, AV_LOG_ERROR, "Error reading lyrics, skipped\n");
432 av_free(descriptor);
433 }
434
435 /**
436 * Parse a comment tag.
437 */
read_comment(AVFormatContext * s,AVIOContext * pb,int taglen,AVDictionary ** metadata)438 static void read_comment(AVFormatContext *s, AVIOContext *pb, int taglen,
439 AVDictionary **metadata)
440 {
441 const char *key = "comment";
442 uint8_t *dst;
443 int encoding, dict_flags = AV_DICT_DONT_OVERWRITE | AV_DICT_DONT_STRDUP_VAL;
444 av_unused int language;
445
446 if (taglen < 4)
447 return;
448
449 encoding = avio_r8(pb);
450 language = avio_rl24(pb);
451 taglen -= 4;
452
453 if (decode_str(s, pb, encoding, &dst, &taglen) < 0) {
454 av_log(s, AV_LOG_ERROR, "Error reading comment frame, skipped\n");
455 return;
456 }
457
458 if (dst && !*dst)
459 av_freep(&dst);
460
461 if (dst) {
462 key = (const char *) dst;
463 dict_flags |= AV_DICT_DONT_STRDUP_KEY;
464 }
465
466 if (decode_str(s, pb, encoding, &dst, &taglen) < 0) {
467 av_log(s, AV_LOG_ERROR, "Error reading comment frame, skipped\n");
468 if (dict_flags & AV_DICT_DONT_STRDUP_KEY)
469 av_freep((void*)&key);
470 return;
471 }
472
473 if (dst)
474 av_dict_set(metadata, key, (const char *) dst, dict_flags);
475 }
476
477 /**
478 * Parse GEOB tag into a ID3v2ExtraMetaGEOB struct.
479 */
read_geobtag(AVFormatContext * s,AVIOContext * pb,int taglen,const char * tag,ID3v2ExtraMeta ** extra_meta,int isv34)480 static void read_geobtag(AVFormatContext *s, AVIOContext *pb, int taglen,
481 const char *tag, ID3v2ExtraMeta **extra_meta,
482 int isv34)
483 {
484 ID3v2ExtraMetaGEOB *geob_data = NULL;
485 ID3v2ExtraMeta *new_extra = NULL;
486 char encoding;
487 unsigned int len;
488
489 if (taglen < 1)
490 return;
491
492 new_extra = av_mallocz(sizeof(ID3v2ExtraMeta));
493 if (!new_extra) {
494 av_log(s, AV_LOG_ERROR, "Failed to alloc %"SIZE_SPECIFIER" bytes\n",
495 sizeof(ID3v2ExtraMeta));
496 return;
497 }
498
499 geob_data = &new_extra->data.geob;
500
501 /* read encoding type byte */
502 encoding = avio_r8(pb);
503 taglen--;
504
505 /* read MIME type (always ISO-8859) */
506 if (decode_str(s, pb, ID3v2_ENCODING_ISO8859, &geob_data->mime_type,
507 &taglen) < 0 ||
508 taglen <= 0)
509 goto fail;
510
511 /* read file name */
512 if (decode_str(s, pb, encoding, &geob_data->file_name, &taglen) < 0 ||
513 taglen <= 0)
514 goto fail;
515
516 /* read content description */
517 if (decode_str(s, pb, encoding, &geob_data->description, &taglen) < 0 ||
518 taglen < 0)
519 goto fail;
520
521 if (taglen) {
522 /* save encapsulated binary data */
523 geob_data->data = av_malloc(taglen);
524 if (!geob_data->data) {
525 av_log(s, AV_LOG_ERROR, "Failed to alloc %d bytes\n", taglen);
526 goto fail;
527 }
528 if ((len = avio_read(pb, geob_data->data, taglen)) < taglen)
529 av_log(s, AV_LOG_WARNING,
530 "Error reading GEOB frame, data truncated.\n");
531 geob_data->datasize = len;
532 } else {
533 geob_data->data = NULL;
534 geob_data->datasize = 0;
535 }
536
537 /* add data to the list */
538 new_extra->tag = "GEOB";
539 new_extra->next = *extra_meta;
540 *extra_meta = new_extra;
541
542 return;
543
544 fail:
545 av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", tag);
546 free_geobtag(geob_data);
547 av_free(new_extra);
548 return;
549 }
550
is_number(const char * str)551 static int is_number(const char *str)
552 {
553 while (*str >= '0' && *str <= '9')
554 str++;
555 return !*str;
556 }
557
get_date_tag(AVDictionary * m,const char * tag)558 static AVDictionaryEntry *get_date_tag(AVDictionary *m, const char *tag)
559 {
560 AVDictionaryEntry *t;
561 if ((t = av_dict_get(m, tag, NULL, AV_DICT_MATCH_CASE)) &&
562 strlen(t->value) == 4 && is_number(t->value))
563 return t;
564 return NULL;
565 }
566
merge_date(AVDictionary ** m)567 static void merge_date(AVDictionary **m)
568 {
569 AVDictionaryEntry *t;
570 char date[17] = { 0 }; // YYYY-MM-DD hh:mm
571
572 if (!(t = get_date_tag(*m, "TYER")) &&
573 !(t = get_date_tag(*m, "TYE")))
574 return;
575 av_strlcpy(date, t->value, 5);
576 av_dict_set(m, "TYER", NULL, 0);
577 av_dict_set(m, "TYE", NULL, 0);
578
579 if (!(t = get_date_tag(*m, "TDAT")) &&
580 !(t = get_date_tag(*m, "TDA")))
581 goto finish;
582 snprintf(date + 4, sizeof(date) - 4, "-%.2s-%.2s", t->value + 2, t->value);
583 av_dict_set(m, "TDAT", NULL, 0);
584 av_dict_set(m, "TDA", NULL, 0);
585
586 if (!(t = get_date_tag(*m, "TIME")) &&
587 !(t = get_date_tag(*m, "TIM")))
588 goto finish;
589 snprintf(date + 10, sizeof(date) - 10,
590 " %.2s:%.2s", t->value, t->value + 2);
591 av_dict_set(m, "TIME", NULL, 0);
592 av_dict_set(m, "TIM", NULL, 0);
593
594 finish:
595 if (date[0])
596 av_dict_set(m, "date", date, 0);
597 }
598
free_apic(void * obj)599 static void free_apic(void *obj)
600 {
601 ID3v2ExtraMetaAPIC *apic = obj;
602 av_buffer_unref(&apic->buf);
603 av_freep(&apic->description);
604 }
605
rstrip_spaces(char * buf)606 static void rstrip_spaces(char *buf)
607 {
608 size_t len = strlen(buf);
609 while (len > 0 && buf[len - 1] == ' ')
610 buf[--len] = 0;
611 }
612
read_apic(AVFormatContext * s,AVIOContext * pb,int taglen,const char * tag,ID3v2ExtraMeta ** extra_meta,int isv34)613 static void read_apic(AVFormatContext *s, AVIOContext *pb, int taglen,
614 const char *tag, ID3v2ExtraMeta **extra_meta,
615 int isv34)
616 {
617 int enc, pic_type;
618 char mimetype[64] = {0};
619 const CodecMime *mime = ff_id3v2_mime_tags;
620 enum AVCodecID id = AV_CODEC_ID_NONE;
621 ID3v2ExtraMetaAPIC *apic = NULL;
622 ID3v2ExtraMeta *new_extra = NULL;
623 int64_t end = avio_tell(pb) + taglen;
624
625 if (taglen <= 4 || (!isv34 && taglen <= 6))
626 goto fail;
627
628 new_extra = av_mallocz(sizeof(*new_extra));
629 if (!new_extra)
630 goto fail;
631
632 apic = &new_extra->data.apic;
633
634 enc = avio_r8(pb);
635 taglen--;
636
637 /* mimetype */
638 if (isv34) {
639 int ret = avio_get_str(pb, taglen, mimetype, sizeof(mimetype));
640 if (ret < 0 || ret >= taglen)
641 goto fail;
642 taglen -= ret;
643 } else {
644 if (avio_read(pb, mimetype, 3) < 0)
645 goto fail;
646
647 mimetype[3] = 0;
648 taglen -= 3;
649 }
650
651 while (mime->id != AV_CODEC_ID_NONE) {
652 if (!av_strncasecmp(mime->str, mimetype, sizeof(mimetype))) {
653 id = mime->id;
654 break;
655 }
656 mime++;
657 }
658 if (id == AV_CODEC_ID_NONE) {
659 av_log(s, AV_LOG_WARNING,
660 "Unknown attached picture mimetype: %s, skipping.\n", mimetype);
661 goto fail;
662 }
663 apic->id = id;
664
665 /* picture type */
666 pic_type = avio_r8(pb);
667 taglen--;
668 if (pic_type < 0 || pic_type >= FF_ARRAY_ELEMS(ff_id3v2_picture_types)) {
669 av_log(s, AV_LOG_WARNING, "Unknown attached picture type %d.\n",
670 pic_type);
671 pic_type = 0;
672 }
673 apic->type = ff_id3v2_picture_types[pic_type];
674
675 /* description and picture data */
676 if (decode_str(s, pb, enc, &apic->description, &taglen) < 0) {
677 av_log(s, AV_LOG_ERROR,
678 "Error decoding attached picture description.\n");
679 goto fail;
680 }
681
682 apic->buf = av_buffer_alloc(taglen + AV_INPUT_BUFFER_PADDING_SIZE);
683 if (!apic->buf || !taglen || avio_read(pb, apic->buf->data, taglen) != taglen)
684 goto fail;
685 memset(apic->buf->data + taglen, 0, AV_INPUT_BUFFER_PADDING_SIZE);
686
687 new_extra->tag = "APIC";
688 new_extra->next = *extra_meta;
689 *extra_meta = new_extra;
690
691 // The description must be unique, and some ID3v2 tag writers add spaces
692 // to write several APIC entries with the same description.
693 rstrip_spaces(apic->description);
694
695 return;
696
697 fail:
698 if (apic)
699 free_apic(apic);
700 av_freep(&new_extra);
701 avio_seek(pb, end, SEEK_SET);
702 }
703
free_chapter(void * obj)704 static void free_chapter(void *obj)
705 {
706 ID3v2ExtraMetaCHAP *chap = obj;
707 av_freep(&chap->element_id);
708 av_dict_free(&chap->meta);
709 }
710
read_chapter(AVFormatContext * s,AVIOContext * pb,int len,const char * ttag,ID3v2ExtraMeta ** extra_meta,int isv34)711 static void read_chapter(AVFormatContext *s, AVIOContext *pb, int len, const char *ttag, ID3v2ExtraMeta **extra_meta, int isv34)
712 {
713 int taglen;
714 char tag[5];
715 ID3v2ExtraMeta *new_extra = NULL;
716 ID3v2ExtraMetaCHAP *chap = NULL;
717
718 new_extra = av_mallocz(sizeof(*new_extra));
719 if (!new_extra)
720 return;
721
722 chap = &new_extra->data.chap;
723
724 if (decode_str(s, pb, 0, &chap->element_id, &len) < 0)
725 goto fail;
726
727 if (len < 16)
728 goto fail;
729
730 chap->start = avio_rb32(pb);
731 chap->end = avio_rb32(pb);
732 avio_skip(pb, 8);
733
734 len -= 16;
735 while (len > 10) {
736 if (avio_read(pb, tag, 4) < 4)
737 goto fail;
738 tag[4] = 0;
739 taglen = avio_rb32(pb);
740 avio_skip(pb, 2);
741 len -= 10;
742 if (taglen < 0 || taglen > len)
743 goto fail;
744 if (tag[0] == 'T')
745 read_ttag(s, pb, taglen, &chap->meta, tag);
746 else
747 avio_skip(pb, taglen);
748 len -= taglen;
749 }
750
751 ff_metadata_conv(&chap->meta, NULL, ff_id3v2_34_metadata_conv);
752 ff_metadata_conv(&chap->meta, NULL, ff_id3v2_4_metadata_conv);
753
754 new_extra->tag = "CHAP";
755 new_extra->next = *extra_meta;
756 *extra_meta = new_extra;
757
758 return;
759
760 fail:
761 free_chapter(chap);
762 av_freep(&new_extra);
763 }
764
free_priv(void * obj)765 static void free_priv(void *obj)
766 {
767 ID3v2ExtraMetaPRIV *priv = obj;
768 av_freep(&priv->owner);
769 av_freep(&priv->data);
770 }
771
read_priv(AVFormatContext * s,AVIOContext * pb,int taglen,const char * tag,ID3v2ExtraMeta ** extra_meta,int isv34)772 static void read_priv(AVFormatContext *s, AVIOContext *pb, int taglen,
773 const char *tag, ID3v2ExtraMeta **extra_meta, int isv34)
774 {
775 ID3v2ExtraMeta *meta;
776 ID3v2ExtraMetaPRIV *priv;
777
778 meta = av_mallocz(sizeof(*meta));
779 if (!meta)
780 return;
781
782 priv = &meta->data.priv;
783
784 if (decode_str(s, pb, ID3v2_ENCODING_ISO8859, &priv->owner, &taglen) < 0)
785 goto fail;
786
787 priv->data = av_malloc(taglen);
788 if (!priv->data)
789 goto fail;
790
791 priv->datasize = taglen;
792
793 if (avio_read(pb, priv->data, priv->datasize) != priv->datasize)
794 goto fail;
795
796 meta->tag = "PRIV";
797 meta->next = *extra_meta;
798 *extra_meta = meta;
799
800 return;
801
802 fail:
803 free_priv(priv);
804 av_freep(&meta);
805 }
806
807 typedef struct ID3v2EMFunc {
808 const char *tag3;
809 const char *tag4;
810 void (*read)(AVFormatContext *s, AVIOContext *pb, int taglen,
811 const char *tag, ID3v2ExtraMeta **extra_meta,
812 int isv34);
813 void (*free)(void *obj);
814 } ID3v2EMFunc;
815
816 static const ID3v2EMFunc id3v2_extra_meta_funcs[] = {
817 { "GEO", "GEOB", read_geobtag, free_geobtag },
818 { "PIC", "APIC", read_apic, free_apic },
819 { "CHAP","CHAP", read_chapter, free_chapter },
820 { "PRIV","PRIV", read_priv, free_priv },
821 { NULL }
822 };
823
824 /**
825 * Get the corresponding ID3v2EMFunc struct for a tag.
826 * @param isv34 Determines if v2.2 or v2.3/4 strings are used
827 * @return A pointer to the ID3v2EMFunc struct if found, NULL otherwise.
828 */
get_extra_meta_func(const char * tag,int isv34)829 static const ID3v2EMFunc *get_extra_meta_func(const char *tag, int isv34)
830 {
831 int i = 0;
832 while (id3v2_extra_meta_funcs[i].tag3) {
833 if (tag && !memcmp(tag,
834 (isv34 ? id3v2_extra_meta_funcs[i].tag4 :
835 id3v2_extra_meta_funcs[i].tag3),
836 (isv34 ? 4 : 3)))
837 return &id3v2_extra_meta_funcs[i];
838 i++;
839 }
840 return NULL;
841 }
842
id3v2_parse(AVIOContext * pb,AVDictionary ** metadata,AVFormatContext * s,int len,uint8_t version,uint8_t flags,ID3v2ExtraMeta ** extra_meta)843 static void id3v2_parse(AVIOContext *pb, AVDictionary **metadata,
844 AVFormatContext *s, int len, uint8_t version,
845 uint8_t flags, ID3v2ExtraMeta **extra_meta)
846 {
847 int isv34, unsync;
848 unsigned tlen;
849 char tag[5];
850 int64_t next, end = avio_tell(pb);
851 int taghdrlen;
852 const char *reason = NULL;
853 AVIOContext pb_local;
854 AVIOContext *pbx;
855 unsigned char *buffer = NULL;
856 int buffer_size = 0;
857 const ID3v2EMFunc *extra_func = NULL;
858 unsigned char *uncompressed_buffer = NULL;
859 av_unused int uncompressed_buffer_size = 0;
860 const char *comm_frame;
861
862 if (end > INT64_MAX - len - 10)
863 return;
864 end += len;
865
866 av_log(s, AV_LOG_DEBUG, "id3v2 ver:%d flags:%02X len:%d\n", version, flags, len);
867
868 switch (version) {
869 case 2:
870 if (flags & 0x40) {
871 reason = "compression";
872 goto error;
873 }
874 isv34 = 0;
875 taghdrlen = 6;
876 comm_frame = "COM";
877 break;
878
879 case 3:
880 case 4:
881 isv34 = 1;
882 taghdrlen = 10;
883 comm_frame = "COMM";
884 break;
885
886 default:
887 reason = "version";
888 goto error;
889 }
890
891 unsync = flags & 0x80;
892
893 if (isv34 && flags & 0x40) { /* Extended header present, just skip over it */
894 int extlen = get_size(pb, 4);
895 if (version == 4)
896 /* In v2.4 the length includes the length field we just read. */
897 extlen -= 4;
898
899 if (extlen < 0) {
900 reason = "invalid extended header length";
901 goto error;
902 }
903 avio_skip(pb, extlen);
904 len -= extlen + 4;
905 if (len < 0) {
906 reason = "extended header too long.";
907 goto error;
908 }
909 }
910
911 while (len >= taghdrlen) {
912 unsigned int tflags = 0;
913 int tunsync = 0;
914 int tcomp = 0;
915 int tencr = 0;
916 unsigned long av_unused dlen;
917
918 if (isv34) {
919 if (avio_read(pb, tag, 4) < 4)
920 break;
921 tag[4] = 0;
922 if (version == 3) {
923 tlen = avio_rb32(pb);
924 } else {
925 /* some encoders incorrectly uses v3 sizes instead of syncsafe ones
926 * so check the next tag to see which one to use */
927 tlen = avio_rb32(pb);
928 if (tlen > 0x7f) {
929 if (tlen < len) {
930 int64_t cur = avio_tell(pb);
931
932 if (ffio_ensure_seekback(pb, 2 /* tflags */ + tlen + 4 /* next tag */))
933 break;
934
935 if (check_tag(pb, cur + 2 + size_to_syncsafe(tlen), 4) == 1)
936 tlen = size_to_syncsafe(tlen);
937 else if (check_tag(pb, cur + 2 + tlen, 4) != 1)
938 break;
939 avio_seek(pb, cur, SEEK_SET);
940 } else
941 tlen = size_to_syncsafe(tlen);
942 }
943 }
944 tflags = avio_rb16(pb);
945 tunsync = tflags & ID3v2_FLAG_UNSYNCH;
946 } else {
947 if (avio_read(pb, tag, 3) < 3)
948 break;
949 tag[3] = 0;
950 tlen = avio_rb24(pb);
951 }
952 if (tlen > (1<<28))
953 break;
954 len -= taghdrlen + tlen;
955
956 if (len < 0)
957 break;
958
959 next = avio_tell(pb) + tlen;
960
961 if (!tlen) {
962 if (tag[0])
963 av_log(s, AV_LOG_DEBUG, "Invalid empty frame %s, skipping.\n",
964 tag);
965 continue;
966 }
967
968 if (tflags & ID3v2_FLAG_DATALEN) {
969 if (tlen < 4)
970 break;
971 dlen = avio_rb32(pb);
972 tlen -= 4;
973 } else
974 dlen = tlen;
975
976 tcomp = tflags & ID3v2_FLAG_COMPRESSION;
977 tencr = tflags & ID3v2_FLAG_ENCRYPTION;
978
979 /* skip encrypted tags and, if no zlib, compressed tags */
980 if (tencr || (!CONFIG_ZLIB && tcomp)) {
981 const char *type;
982 if (!tcomp)
983 type = "encrypted";
984 else if (!tencr)
985 type = "compressed";
986 else
987 type = "encrypted and compressed";
988
989 av_log(s, AV_LOG_WARNING, "Skipping %s ID3v2 frame %s.\n", type, tag);
990 avio_skip(pb, tlen);
991 /* check for text tag or supported special meta tag */
992 } else if (tag[0] == 'T' ||
993 !memcmp(tag, "USLT", 4) ||
994 !strcmp(tag, comm_frame) ||
995 (extra_meta &&
996 (extra_func = get_extra_meta_func(tag, isv34)))) {
997 pbx = pb;
998
999 if (unsync || tunsync || tcomp) {
1000 av_fast_malloc(&buffer, &buffer_size, tlen);
1001 if (!buffer) {
1002 av_log(s, AV_LOG_ERROR, "Failed to alloc %d bytes\n", tlen);
1003 goto seek;
1004 }
1005 }
1006 if (unsync || tunsync) {
1007 uint8_t *b = buffer;
1008 uint8_t *t = buffer;
1009 uint8_t *end = t + tlen;
1010
1011 if (avio_read(pb, buffer, tlen) != tlen) {
1012 av_log(s, AV_LOG_ERROR, "Failed to read tag data\n");
1013 goto seek;
1014 }
1015
1016 while (t != end) {
1017 *b++ = *t++;
1018 if (t != end && t[-1] == 0xff && !t[0])
1019 t++;
1020 }
1021
1022 ffio_init_context(&pb_local, buffer, b - buffer, 0, NULL, NULL, NULL,
1023 NULL);
1024 tlen = b - buffer;
1025 pbx = &pb_local; // read from sync buffer
1026 }
1027
1028 #if CONFIG_ZLIB
1029 if (tcomp) {
1030 int err;
1031
1032 av_log(s, AV_LOG_DEBUG, "Compresssed frame %s tlen=%d dlen=%ld\n", tag, tlen, dlen);
1033
1034 if (tlen <= 0)
1035 goto seek;
1036 if (dlen / 32768 > tlen)
1037 goto seek;
1038
1039 av_fast_malloc(&uncompressed_buffer, &uncompressed_buffer_size, dlen);
1040 if (!uncompressed_buffer) {
1041 av_log(s, AV_LOG_ERROR, "Failed to alloc %ld bytes\n", dlen);
1042 goto seek;
1043 }
1044
1045 if (!(unsync || tunsync)) {
1046 err = avio_read(pb, buffer, tlen);
1047 if (err < 0) {
1048 av_log(s, AV_LOG_ERROR, "Failed to read compressed tag\n");
1049 goto seek;
1050 }
1051 tlen = err;
1052 }
1053
1054 err = uncompress(uncompressed_buffer, &dlen, buffer, tlen);
1055 if (err != Z_OK) {
1056 av_log(s, AV_LOG_ERROR, "Failed to uncompress tag: %d\n", err);
1057 goto seek;
1058 }
1059 ffio_init_context(&pb_local, uncompressed_buffer, dlen, 0, NULL, NULL, NULL, NULL);
1060 tlen = dlen;
1061 pbx = &pb_local; // read from sync buffer
1062 }
1063 #endif
1064 if (tag[0] == 'T')
1065 /* parse text tag */
1066 read_ttag(s, pbx, tlen, metadata, tag);
1067 else if (!memcmp(tag, "USLT", 4))
1068 read_uslt(s, pbx, tlen, metadata);
1069 else if (!strcmp(tag, comm_frame))
1070 read_comment(s, pbx, tlen, metadata);
1071 else
1072 /* parse special meta tag */
1073 extra_func->read(s, pbx, tlen, tag, extra_meta, isv34);
1074 } else if (!tag[0]) {
1075 if (tag[1])
1076 av_log(s, AV_LOG_WARNING, "invalid frame id, assuming padding\n");
1077 avio_skip(pb, tlen);
1078 break;
1079 }
1080 /* Skip to end of tag */
1081 seek:
1082 avio_seek(pb, next, SEEK_SET);
1083 }
1084
1085 /* Footer preset, always 10 bytes, skip over it */
1086 if (version == 4 && flags & 0x10)
1087 end += 10;
1088
1089 error:
1090 if (reason)
1091 av_log(s, AV_LOG_INFO, "ID3v2.%d tag skipped, cannot handle %s\n",
1092 version, reason);
1093 avio_seek(pb, end, SEEK_SET);
1094 av_free(buffer);
1095 av_free(uncompressed_buffer);
1096 return;
1097 }
1098
id3v2_read_internal(AVIOContext * pb,AVDictionary ** metadata,AVFormatContext * s,const char * magic,ID3v2ExtraMeta ** extra_meta,int64_t max_search_size)1099 static void id3v2_read_internal(AVIOContext *pb, AVDictionary **metadata,
1100 AVFormatContext *s, const char *magic,
1101 ID3v2ExtraMeta **extra_meta, int64_t max_search_size)
1102 {
1103 int len, ret;
1104 uint8_t buf[ID3v2_HEADER_SIZE];
1105 int found_header;
1106 int64_t start, off;
1107
1108 if (max_search_size && max_search_size < ID3v2_HEADER_SIZE)
1109 return;
1110
1111 start = avio_tell(pb);
1112 do {
1113 /* save the current offset in case there's nothing to read/skip */
1114 off = avio_tell(pb);
1115 if (max_search_size && off - start >= max_search_size - ID3v2_HEADER_SIZE) {
1116 avio_seek(pb, off, SEEK_SET);
1117 break;
1118 }
1119
1120 ret = ffio_ensure_seekback(pb, ID3v2_HEADER_SIZE);
1121 if (ret >= 0)
1122 ret = avio_read(pb, buf, ID3v2_HEADER_SIZE);
1123 if (ret != ID3v2_HEADER_SIZE) {
1124 avio_seek(pb, off, SEEK_SET);
1125 break;
1126 }
1127 found_header = ff_id3v2_match(buf, magic);
1128 if (found_header) {
1129 /* parse ID3v2 header */
1130 len = ((buf[6] & 0x7f) << 21) |
1131 ((buf[7] & 0x7f) << 14) |
1132 ((buf[8] & 0x7f) << 7) |
1133 (buf[9] & 0x7f);
1134 id3v2_parse(pb, metadata, s, len, buf[3], buf[5], extra_meta);
1135 } else {
1136 avio_seek(pb, off, SEEK_SET);
1137 }
1138 } while (found_header);
1139 ff_metadata_conv(metadata, NULL, ff_id3v2_34_metadata_conv);
1140 ff_metadata_conv(metadata, NULL, id3v2_2_metadata_conv);
1141 ff_metadata_conv(metadata, NULL, ff_id3v2_4_metadata_conv);
1142 merge_date(metadata);
1143 }
1144
ff_id3v2_read_dict(AVIOContext * pb,AVDictionary ** metadata,const char * magic,ID3v2ExtraMeta ** extra_meta)1145 void ff_id3v2_read_dict(AVIOContext *pb, AVDictionary **metadata,
1146 const char *magic, ID3v2ExtraMeta **extra_meta)
1147 {
1148 id3v2_read_internal(pb, metadata, NULL, magic, extra_meta, 0);
1149 }
1150
ff_id3v2_read(AVFormatContext * s,const char * magic,ID3v2ExtraMeta ** extra_meta,unsigned int max_search_size)1151 void ff_id3v2_read(AVFormatContext *s, const char *magic,
1152 ID3v2ExtraMeta **extra_meta, unsigned int max_search_size)
1153 {
1154 id3v2_read_internal(s->pb, &s->metadata, s, magic, extra_meta, max_search_size);
1155 }
1156
ff_id3v2_free_extra_meta(ID3v2ExtraMeta ** extra_meta)1157 void ff_id3v2_free_extra_meta(ID3v2ExtraMeta **extra_meta)
1158 {
1159 ID3v2ExtraMeta *current = *extra_meta, *next;
1160 const ID3v2EMFunc *extra_func;
1161
1162 while (current) {
1163 if ((extra_func = get_extra_meta_func(current->tag, 1)))
1164 extra_func->free(¤t->data);
1165 next = current->next;
1166 av_freep(¤t);
1167 current = next;
1168 }
1169
1170 *extra_meta = NULL;
1171 }
1172
ff_id3v2_parse_apic(AVFormatContext * s,ID3v2ExtraMeta * extra_meta)1173 int ff_id3v2_parse_apic(AVFormatContext *s, ID3v2ExtraMeta *extra_meta)
1174 {
1175 ID3v2ExtraMeta *cur;
1176
1177 for (cur = extra_meta; cur; cur = cur->next) {
1178 ID3v2ExtraMetaAPIC *apic;
1179 AVStream *st;
1180
1181 if (strcmp(cur->tag, "APIC"))
1182 continue;
1183 apic = &cur->data.apic;
1184
1185 if (!(st = avformat_new_stream(s, NULL)))
1186 return AVERROR(ENOMEM);
1187
1188 st->disposition |= AV_DISPOSITION_ATTACHED_PIC;
1189 st->codecpar->codec_type = AVMEDIA_TYPE_VIDEO;
1190 st->codecpar->codec_id = apic->id;
1191
1192 if (AV_RB64(apic->buf->data) == PNGSIG)
1193 st->codecpar->codec_id = AV_CODEC_ID_PNG;
1194
1195 if (apic->description[0])
1196 av_dict_set(&st->metadata, "title", apic->description, 0);
1197
1198 av_dict_set(&st->metadata, "comment", apic->type, 0);
1199
1200 av_packet_unref(&st->attached_pic);
1201 st->attached_pic.buf = apic->buf;
1202 st->attached_pic.data = apic->buf->data;
1203 st->attached_pic.size = apic->buf->size - AV_INPUT_BUFFER_PADDING_SIZE;
1204 st->attached_pic.stream_index = st->index;
1205 st->attached_pic.flags |= AV_PKT_FLAG_KEY;
1206
1207 apic->buf = NULL;
1208 }
1209
1210 return 0;
1211 }
1212
ff_id3v2_parse_chapters(AVFormatContext * s,ID3v2ExtraMeta * extra_meta)1213 int ff_id3v2_parse_chapters(AVFormatContext *s, ID3v2ExtraMeta *extra_meta)
1214 {
1215 int ret = 0;
1216 ID3v2ExtraMeta *cur;
1217 AVRational time_base = {1, 1000};
1218 ID3v2ExtraMetaCHAP **chapters = NULL;
1219 int num_chapters = 0;
1220 int i;
1221
1222 // since extra_meta is a linked list where elements are prepended,
1223 // we need to reverse the order of chapters
1224 for (cur = extra_meta; cur; cur = cur->next) {
1225 ID3v2ExtraMetaCHAP *chap;
1226
1227 if (strcmp(cur->tag, "CHAP"))
1228 continue;
1229 chap = &cur->data.chap;
1230
1231 if ((ret = av_dynarray_add_nofree(&chapters, &num_chapters, chap)) < 0)
1232 goto end;
1233 }
1234
1235 for (i = 0; i < (num_chapters / 2); i++) {
1236 ID3v2ExtraMetaCHAP *right;
1237 int right_index;
1238
1239 right_index = (num_chapters - 1) - i;
1240 right = chapters[right_index];
1241
1242 chapters[right_index] = chapters[i];
1243 chapters[i] = right;
1244 }
1245
1246 for (i = 0; i < num_chapters; i++) {
1247 ID3v2ExtraMetaCHAP *chap;
1248 AVChapter *chapter;
1249
1250 chap = chapters[i];
1251 chapter = avpriv_new_chapter(s, i, time_base, chap->start, chap->end, chap->element_id);
1252 if (!chapter)
1253 continue;
1254
1255 if ((ret = av_dict_copy(&chapter->metadata, chap->meta, 0)) < 0)
1256 goto end;
1257 }
1258
1259 end:
1260 av_freep(&chapters);
1261 return ret;
1262 }
1263
ff_id3v2_parse_priv_dict(AVDictionary ** metadata,ID3v2ExtraMeta * extra_meta)1264 int ff_id3v2_parse_priv_dict(AVDictionary **metadata, ID3v2ExtraMeta *extra_meta)
1265 {
1266 ID3v2ExtraMeta *cur;
1267 int dict_flags = AV_DICT_DONT_OVERWRITE | AV_DICT_DONT_STRDUP_KEY | AV_DICT_DONT_STRDUP_VAL;
1268
1269 for (cur = extra_meta; cur; cur = cur->next) {
1270 if (!strcmp(cur->tag, "PRIV")) {
1271 ID3v2ExtraMetaPRIV *priv = &cur->data.priv;
1272 AVBPrint bprint;
1273 char *escaped, *key;
1274 int i, ret;
1275
1276 if ((key = av_asprintf(ID3v2_PRIV_METADATA_PREFIX "%s", priv->owner)) == NULL) {
1277 return AVERROR(ENOMEM);
1278 }
1279
1280 av_bprint_init(&bprint, priv->datasize + 1, AV_BPRINT_SIZE_UNLIMITED);
1281
1282 for (i = 0; i < priv->datasize; i++) {
1283 if (priv->data[i] < 32 || priv->data[i] > 126 || priv->data[i] == '\\') {
1284 av_bprintf(&bprint, "\\x%02x", priv->data[i]);
1285 } else {
1286 av_bprint_chars(&bprint, priv->data[i], 1);
1287 }
1288 }
1289
1290 if ((ret = av_bprint_finalize(&bprint, &escaped)) < 0) {
1291 av_free(key);
1292 return ret;
1293 }
1294
1295 if ((ret = av_dict_set(metadata, key, escaped, dict_flags)) < 0) {
1296 return ret;
1297 }
1298 }
1299 }
1300
1301 return 0;
1302 }
1303
ff_id3v2_parse_priv(AVFormatContext * s,ID3v2ExtraMeta * extra_meta)1304 int ff_id3v2_parse_priv(AVFormatContext *s, ID3v2ExtraMeta *extra_meta)
1305 {
1306 return ff_id3v2_parse_priv_dict(&s->metadata, extra_meta);
1307 }
1308