2 * Copyright (C) 2008 by INdT
4 * This program is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU Lesser General Public License
6 * as published by the Free Software Foundation; either version 2
7 * of the License, or (at your option) any later version.
9 * This program is distributed in the hope that it will be useful,
10 * but WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 * GNU General Public License for more details.
14 * You should have received a copy of the GNU Lesser General Public License
15 * along with this program; if not, write to the Free Software
16 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
18 * @author Andre Moreira Magalhaes <andre.magalhaes@openbossa.org>
24 * asf/wma file parser.
27 #include <lightmediascanner_plugin.h>
28 #include <lightmediascanner_db.h>
31 #include <sys/types.h>
39 #define NSEC100_PER_SEC 10000000ULL
40 #define MSEC_PER_SEC 1000ULL
43 STREAM_TYPE_UNKNOWN = 0,
49 ATTR_TYPE_UNICODE = 0,
59 struct lms_string_size title;
60 struct lms_string_size artist;
61 struct lms_string_size album;
62 struct lms_string_size genre;
63 unsigned char trackno;
67 struct lms_plugin plugin;
68 lms_db_audio_t *audio_db;
69 lms_db_video_t *video_db;
70 lms_charset_conv_t *cs_conv;
73 static const char _name[] = "asf";
74 static const struct lms_string_size _container = LMS_STATIC_STRING_SIZE("asf");
75 static const struct lms_string_size _exts[] = {
76 LMS_STATIC_STRING_SIZE(".wma"),
77 LMS_STATIC_STRING_SIZE(".wmv"),
78 LMS_STATIC_STRING_SIZE(".asf")
80 static const char *_cats[] = {
85 static const char *_authors[] = {
86 "Andre Moreira Magalhaes",
92 enum StreamTypes type;
98 uint32_t sampling_rate;
104 /* TODO: Add the gazillion of possible codecs -- possibly a task to gperf */
105 static const struct {
107 struct lms_string_size name;
109 { 0x0160, LMS_STATIC_STRING_SIZE("wmav1") },
110 { 0x0161, LMS_STATIC_STRING_SIZE("wmav2") },
111 { 0x0162, LMS_STATIC_STRING_SIZE("wmavpro") },
112 { 0x0163, LMS_STATIC_STRING_SIZE("wmavlossless") },
113 { 0x1600, LMS_STATIC_STRING_SIZE("aac") },
114 { 0x706d, LMS_STATIC_STRING_SIZE("aac") },
115 { 0x4143, LMS_STATIC_STRING_SIZE("aac") },
116 { 0xA106, LMS_STATIC_STRING_SIZE("aac") },
117 { 0xF1AC, LMS_STATIC_STRING_SIZE("flac") },
118 { 0x0055, LMS_STATIC_STRING_SIZE("mp3") },
124 * Microsoft defines these 16-byte (128-bit) GUIDs as:
125 * first 8 bytes are in little-endian order
126 * next 8 bytes are in big-endian order
128 * Eg.: AaBbCcDd-EeFf-GgHh-IiJj-KkLlMmNnOoPp:
130 * to convert to byte string do as follow:
132 * $Dd $Cc $Bb $Aa $Ff $Ee $Hh $Gg $Ii $Jj $Kk $Ll $Mm $Nn $Oo $Pp
134 * See http://www.microsoft.com/windows/windowsmedia/forpros/format/asfspec.aspx
136 static const char header_guid[16] = "\x30\x26\xB2\x75\x8E\x66\xCF\x11\xA6\xD9\x00\xAA\x00\x62\xCE\x6C";
137 static const char file_properties_guid[16] = "\xA1\xDC\xAB\x8C\x47\xA9\xCF\x11\x8E\xE4\x00\xC0\x0C\x20\x53\x65";
138 static const char stream_properties_guid[16] = "\x91\x07\xDC\xB7\xB7\xA9\xCF\x11\x8E\xE6\x00\xC0\x0C\x20\x53\x65";
139 static const char stream_type_audio_guid[16] = "\x40\x9E\x69\xF8\x4D\x5B\xCF\x11\xA8\xFD\x00\x80\x5F\x5C\x44\x2B";
140 static const char stream_type_video_guid[16] = "\xC0\xEF\x19\xBC\x4D\x5B\xCF\x11\xA8\xFD\x00\x80\x5F\x5C\x44\x2B";
141 static const char content_description_guid[16] = "\x33\x26\xB2\x75\x8E\x66\xCF\x11\xA6\xD9\x00\xAA\x00\x62\xCE\x6C";
142 static const char extended_content_description_guid[16] = "\x40\xA4\xD0\xD2\x07\xE3\xD2\x11\x97\xF0\x00\xA0\xC9\x5E\xA8\x50";
143 static const char header_extension_guid[16] = "\xb5\x03\xbf_.\xa9\xcf\x11\x8e\xe3\x00\xc0\x0c Se";
144 static const char metadata_guid[16] = "\xEA\xCB\xF8\xC5\xAF[wH\204g\xAA\214D\xFAL\xCA";
145 static const char metadata_library_guid[16] = "\224\034#D\230\224\321I\241A\x1d\x13NEpT";
146 static const char content_encryption_object_guid[16] = "\xFB\xB3\x11\x22\x23\xBD\xD2\x11\xB4\xB7\x00\xA0\xC9\x55\xFC\x6E";
147 static const char extended_content_encryption_object_guid[16] = "\x14\xE6\x8A\x29\x22\x26\x17\x4C\xB9\x35\xDA\xE0\x7E\xE9\x28\x9C";
149 static const char attr_name_wm_album_artist[28] = "\x57\x00\x4d\x00\x2f\x00\x41\x00\x6c\x00\x62\x00\x75\x00\x6d\x00\x41\x00\x72\x00\x74\x00\x69\x00\x73\x00\x74\x00";
150 static const char attr_name_wm_album_title[26] = "\x57\x00\x4d\x00\x2f\x00\x41\x00\x6c\x00\x62\x00\x75\x00\x6d\x00\x54\x00\x69\x00\x74\x00\x6c\x00\x65\x00";
151 static const char attr_name_wm_genre[16] = "\x57\x00\x4d\x00\x2f\x00\x47\x00\x65\x00\x6e\x00\x72\x00\x65\x00";
152 static const char attr_name_wm_track_number[28] = "\x57\x00\x4d\x00\x2f\x00\x54\x00\x72\x00\x61\x00\x63\x00\x6b\x00\x4e\x00\x75\x00\x6d\x00\x62\x00\x65\x00\x72\x00";
155 _to_number(const char *data, unsigned int type_size, unsigned int data_size)
158 unsigned int last, i;
160 last = data_size > type_size ? type_size : data_size;
162 for (i = 0; i < last; i++)
163 sum |= (unsigned char) (data[i]) << (i * 8);
172 if (read(fd, &v, 2) != 2)
174 return (short) _to_number(v, sizeof(unsigned short), 2);
181 if (read(fd, &v, 4) != 4)
183 return (unsigned int) _to_number(v, sizeof(unsigned int), 4);
190 if (read(fd, &v, 8) != 8)
192 return _to_number(v, sizeof(unsigned long long), 8);
196 _read_string(int fd, size_t count, char **str, unsigned int *len)
199 ssize_t data_size, size;
201 data = malloc(sizeof(char) * count);
202 data_size = read(fd, data, count);
203 if (data_size == -1) {
210 if (data[size - 1] != '\0' || data[size - 2] != '\0')
222 _parse_file_properties(lms_charset_conv_t *cs_conv, int fd,
223 struct lms_audio_info *info)
228 uint64_t creation_date;
229 uint64_t data_packets_count;
230 uint64_t play_duration;
231 uint64_t send_duration;
234 uint32_t min_data_packet_size;
235 uint32_t max_data_packet_size;
236 uint32_t max_bitrate;
237 } __attribute__((packed)) props;
240 r = read(fd, &props, sizeof(props));
241 if (r != sizeof(props))
245 if (le32toh(props.flags) & 0x1)
248 /* ASF spec 01.20.06 sec. 3.2: we need to subtract the preroll value from
249 * the duration in order to obtain the real duration */
250 info->length = (unsigned int)((le64toh(props.play_duration) / NSEC100_PER_SEC)
251 - le64toh(props.preroll) / MSEC_PER_SEC);
256 static struct lms_string_size
257 _codec_id_to_str(uint16_t id)
261 for (i = 0; _codecs[i].name.str != NULL; i++)
262 if (_codecs[i].id == id)
263 return _codecs[i].name;
265 return _codecs[i].name;
269 _parse_stream_properties(int fd, struct stream **pstream)
272 char stream_type[16];
273 char error_correction_type[16];
274 uint64_t time_offset;
275 uint32_t type_specific_len;
276 uint32_t error_correction_data_len;
278 uint32_t reserved; /* don't use, unaligned */
279 } __attribute__((packed)) props;
286 s = calloc(1, sizeof(struct stream));
290 r = read(fd, &props, sizeof(props));
291 if (r != sizeof(props))
294 if (memcmp(props.stream_type, stream_type_audio_guid, 16) == 0)
295 s->type = STREAM_TYPE_AUDIO;
296 else if (memcmp(props.stream_type, stream_type_video_guid, 16) == 0)
297 s->type = STREAM_TYPE_VIDEO;
303 s->id = le16toh(props.flags) & 0x7F;
304 /* Not a valid stream */
308 if (s->type == STREAM_TYPE_AUDIO) {
309 if (le32toh(props.type_specific_len) < 18)
312 s->audio.codec_id = _read_word(fd);
313 s->audio.channels = _read_word(fd);
314 s->audio.sampling_rate = _read_dword(fd);
315 s->audio.byterate = _read_dword(fd);
329 _parse_content_description(lms_charset_conv_t *cs_conv, int fd, struct asf_info *info)
331 int title_length = _read_word(fd);
332 int artist_length = _read_word(fd);
333 int copyright_length = _read_word(fd);
334 int comment_length = _read_word(fd);
335 int rating_length = _read_word(fd);
337 _read_string(fd, title_length, &info->title.str, &info->title.len);
338 lms_charset_conv_force(cs_conv, &info->title.str, &info->title.len);
339 _read_string(fd, artist_length, &info->artist.str, &info->artist.len);
340 lms_charset_conv_force(cs_conv, &info->artist.str, &info->artist.len);
341 /* ignore copyright, comment and rating */
342 lseek(fd, copyright_length + comment_length + rating_length, SEEK_CUR);
346 _parse_attribute_name(int fd,
348 unsigned int *attr_name_len,
352 int attr_name_length;
354 attr_name_length = _read_word(fd);
355 _read_string(fd, attr_name_length, attr_name, attr_name_len);
356 *attr_type = _read_word(fd);
357 *attr_size = _read_word(fd);
361 _parse_attribute_string_data(lms_charset_conv_t *cs_conv,
365 unsigned int *attr_data_len)
367 _read_string(fd, attr_size, attr_data, attr_data_len);
368 lms_charset_conv_force(cs_conv, attr_data, attr_data_len);
372 _skip_attribute_data(int fd, int kind, int attr_type, int attr_size)
376 lseek(fd, 2, SEEK_CUR);
381 lseek(fd, 4, SEEK_CUR);
383 lseek(fd, 2, SEEK_CUR);
386 case ATTR_TYPE_DWORD:
387 lseek(fd, 4, SEEK_CUR);
390 case ATTR_TYPE_QWORD:
391 lseek(fd, 8, SEEK_CUR);
394 case ATTR_TYPE_UNICODE:
395 case ATTR_TYPE_BYTES:
397 lseek(fd, attr_size, SEEK_CUR);
406 _parse_extended_content_description_object(lms_charset_conv_t *cs_conv, int fd, struct asf_info *info)
408 int count = _read_word(fd);
410 unsigned int attr_name_len;
411 int attr_type, attr_size;
414 _parse_attribute_name(fd,
415 &attr_name, &attr_name_len,
416 &attr_type, &attr_size);
417 if (attr_type == ATTR_TYPE_UNICODE) {
418 if (memcmp(attr_name, attr_name_wm_album_title, attr_name_len) == 0)
419 _parse_attribute_string_data(cs_conv,
423 else if (memcmp(attr_name, attr_name_wm_genre, attr_name_len) == 0)
424 _parse_attribute_string_data(cs_conv,
428 else if (memcmp(attr_name, attr_name_wm_album_artist, attr_name_len) == 0)
429 _parse_attribute_string_data(cs_conv,
433 else if (memcmp(attr_name, attr_name_wm_track_number, attr_name_len) == 0) {
435 unsigned int trackno_len;
436 _parse_attribute_string_data(cs_conv,
441 info->trackno = atoi(trackno);
446 _skip_attribute_data(fd, 0, attr_type, attr_size);
449 _skip_attribute_data(fd, 0, attr_type, attr_size);
455 _match(struct plugin *p, const char *path, int len, int base)
459 i = lms_which_extension(path, len, _exts, LMS_ARRAY_SIZE(_exts));
463 return (void*)(i + 1);
467 _parse(struct plugin *plugin, struct lms_context *ctxt, const struct lms_file_info *finfo, void *match)
469 struct asf_info info = { };
470 struct lms_audio_info audio_info = { };
471 struct lms_video_info video_info = { };
472 int r, fd, num_objects, i;
475 int stream_type = STREAM_TYPE_UNKNOWN;
476 struct stream *streams = NULL;
478 fd = open(finfo->path, O_RDONLY);
484 if (read(fd, &guid, 16) != 16) {
489 if (memcmp(guid, header_guid, 16) != 0) {
490 fprintf(stderr, "ERROR: invalid header (%s).\n", finfo->path);
495 size = _read_qword(fd);
496 num_objects = _read_dword(fd);
498 lseek(fd, 2, SEEK_CUR);
500 for (i = 0; i < num_objects; ++i) {
502 size = _read_qword(fd);
504 if (memcmp(guid, file_properties_guid, 16) == 0) {
505 r = _parse_file_properties(plugin->cs_conv, fd, &audio_info);
508 lseek(fd, size - (24 + r), SEEK_CUR);
509 } else if (memcmp(guid, stream_properties_guid, 16) == 0) {
511 r = _parse_stream_properties(fd, &s);
515 lseek(fd, size - (24 + r), SEEK_CUR);
519 if (stream_type != STREAM_TYPE_VIDEO)
520 stream_type = s->type;
524 } else if (memcmp(guid, content_description_guid, 16) == 0)
525 _parse_content_description(plugin->cs_conv, fd, &info);
526 else if (memcmp(guid, extended_content_description_guid, 16) == 0)
527 _parse_extended_content_description_object(plugin->cs_conv, fd, &info);
528 else if (memcmp(guid, content_encryption_object_guid, 16) == 0 ||
529 memcmp(guid, extended_content_encryption_object_guid, 16) == 0) {
530 /* ignore DRM'd files */
531 fprintf(stderr, "ERROR: ignoring DRM'd file %s\n", finfo->path);
535 lseek(fd, size - 24, SEEK_CUR);
538 /* try to define stream type by extension */
539 if (stream_type == STREAM_TYPE_UNKNOWN) {
540 long ext_idx = ((long)match) - 1;
541 if (strcmp(_exts[ext_idx].str, ".wma") == 0)
542 stream_type = STREAM_TYPE_AUDIO;
543 /* consider wmv and asf as video */
545 stream_type = STREAM_TYPE_VIDEO;
548 lms_string_size_strip_and_free(&info.title);
549 lms_string_size_strip_and_free(&info.artist);
550 lms_string_size_strip_and_free(&info.album);
551 lms_string_size_strip_and_free(&info.genre);
553 if (!info.title.str) {
555 ext_idx = ((long)match) - 1;
556 info.title.len = finfo->path_len - finfo->base - _exts[ext_idx].len;
557 info.title.str = malloc((info.title.len + 1) * sizeof(char));
558 memcpy(info.title.str, finfo->path + finfo->base, info.title.len);
559 info.title.str[info.title.len] = '\0';
560 lms_charset_conv(ctxt->cs_conv, &info.title.str, &info.title.len);
564 fprintf(stderr, "file %s info\n", finfo->path);
565 fprintf(stderr, "\ttitle='%s'\n", info.title.str);
566 fprintf(stderr, "\tartist='%s'\n", info.artist.str);
567 fprintf(stderr, "\talbum='%s'\n", info.album.str);
568 fprintf(stderr, "\tgenre='%s'\n", info.genre.str);
569 fprintf(stderr, "\ttrackno=%d\n", info.trackno);
572 audio_info.container = _container;
574 if (stream_type == STREAM_TYPE_AUDIO) {
575 audio_info.id = finfo->id;
576 audio_info.title = info.title;
577 audio_info.artist = info.artist;
578 audio_info.album = info.album;
579 audio_info.genre = info.genre;
580 audio_info.trackno = info.trackno;
582 audio_info.channels = streams->audio.channels;
583 audio_info.bitrate = streams->audio.byterate * 8;
584 audio_info.sampling_rate = streams->audio.sampling_rate;
585 audio_info.codec = _codec_id_to_str(streams->audio.codec_id);
587 r = lms_db_audio_add(plugin->audio_db, &audio_info);
589 video_info.id = finfo->id;
590 video_info.title = info.title;
591 video_info.artist = info.artist;
592 r = lms_db_video_add(plugin->video_db, &video_info);
597 struct stream *s = streams;
602 free(info.title.str);
603 free(info.artist.str);
604 free(info.album.str);
605 free(info.genre.str);
607 posix_fadvise(fd, 0, 0, POSIX_FADV_DONTNEED);
614 _setup(struct plugin *plugin, struct lms_context *ctxt)
616 plugin->audio_db = lms_db_audio_new(ctxt->db);
617 if (!plugin->audio_db)
619 plugin->video_db = lms_db_video_new(ctxt->db);
620 if (!plugin->video_db)
622 plugin->cs_conv = lms_charset_conv_new();
623 if (!plugin->cs_conv)
625 lms_charset_conv_add(plugin->cs_conv, "UTF-16LE");
631 _start(struct plugin *plugin, struct lms_context *ctxt)
634 r = lms_db_audio_start(plugin->audio_db);
635 r |= lms_db_video_start(plugin->video_db);
640 _finish(struct plugin *plugin, struct lms_context *ctxt)
642 if (plugin->audio_db)
643 lms_db_audio_free(plugin->audio_db);
644 if (plugin->video_db)
645 lms_db_video_free(plugin->video_db);
647 lms_charset_conv_free(plugin->cs_conv);
653 _close(struct plugin *plugin)
659 API struct lms_plugin *
660 lms_plugin_open(void)
662 struct plugin *plugin;
664 plugin = (struct plugin *)malloc(sizeof(*plugin));
665 plugin->plugin.name = _name;
666 plugin->plugin.match = (lms_plugin_match_fn_t)_match;
667 plugin->plugin.parse = (lms_plugin_parse_fn_t)_parse;
668 plugin->plugin.close = (lms_plugin_close_fn_t)_close;
669 plugin->plugin.setup = (lms_plugin_setup_fn_t)_setup;
670 plugin->plugin.start = (lms_plugin_start_fn_t)_start;
671 plugin->plugin.finish = (lms_plugin_finish_fn_t)_finish;
673 return (struct lms_plugin *)plugin;
676 API const struct lms_plugin_info *
677 lms_plugin_info(void)
679 static struct lms_plugin_info info = {
682 "Microsoft WMA, WMV and ASF",
685 "http://lms.garage.maemo.org"