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