Nyaaori/rockbox
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

Install the single stage synchronous audio initialization I cooked up awhile back.

Browse source 
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@12693 a1c6a512-1295-4272-9138-f99709370657
This commit is contained in:
Michael Sevakis 2007-03-09 12:30:14 +00:00
parent 8e7ab3dfff
commit 98cd3e8868
4 changed files with 128 additions and 179 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

21
apps/main.c
View file

@ -244,9 +244,6 @@ static void init(void)
button_init();
backlight_init();
lang_init();
#if CONFIG_CODEC == SWCODEC
audio_preinit(); /* Must be done before settings_apply() */
#endif
/* Must be done before any code uses the multi-screen APi */
screen_access_init();
gui_syncstatusbar_init(&statusbars);
@ -262,6 +259,8 @@ static void init(void)
sleep(HZ/2);
tree_init();
playlist_init();
#if CONFIG_CODEC != SWCODEC
mp3_init( global_settings.volume,
global_settings.bass,
global_settings.treble,
@ -277,14 +276,13 @@ static void init(void)
global_settings.mdb_enable,
global_settings.superbass);
/* audio_init must to know the size of voice buffer so init voice first */
talk_init();
#endif /* CONFIG_CODEC != SWCODEC */
scrobbler_init();
cuesheet_init();
/* audio_init must to know the size of voice buffer so init voice first */
#if CONFIG_CODEC == SWCODEC
talk_init();
#endif
audio_init();
button_clear_queue(); /* Empty the keyboard buffer */
}
@ -365,10 +363,6 @@ static void init(void)
powermgmt_init();
#if CONFIG_CODEC == SWCODEC
audio_preinit();
#endif
#if CONFIG_TUNER
radio_init();
#endif
@ -499,6 +493,7 @@ static void init(void)
scrobbler_init();
cuesheet_init();
#if CONFIG_CODEC != SWCODEC
/* No buffer allocation (see buffer.c) may take place after the call to
audio_init() since the mpeg thread takes the rest of the buffer space */
mp3_init( global_settings.volume,
@ -518,8 +513,10 @@ static void init(void)
/* audio_init must to know the size of voice buffer so init voice first */
talk_init();
#endif /* CONFIG_CODEC != SWCODEC */
audio_init();
#if (CONFIG_CODEC == SWCODEC) && defined(HAVE_RECORDING) && !defined(SIMULATOR)
pcm_rec_init();
#endif

278
apps/playback.c
View file

@ -832,57 +832,11 @@ void audio_set_crossfade(int enable)
audio_play(offset);
}
void audio_preinit(void)
{
logf("playback system pre-init");
filling = false;
set_current_codec(CODEC_IDX_AUDIO);
playing = false;
paused = false;
audio_codec_loaded = false;
#ifdef PLAYBACK_VOICE
voice_is_playing = false;
voice_thread_start = false;
voice_codec_loaded = false;
#endif
track_changed = false;
current_fd = -1;
track_buffer_callback = NULL;
track_unbuffer_callback = NULL;
track_changed_callback = NULL;
track_ridx = 0; /* Just to prevent CUR_TI from being anything random. */
prev_ti = &tracks[MAX_TRACK-1]; /* And prevent prev_ti being random too */
#ifdef PLAYBACK_VOICE
mutex_init(&mutex_codecthread);
#endif
queue_init(&audio_queue, true);
queue_enable_queue_send(&audio_queue, &audio_queue_sender_list);
queue_init(&codec_queue, true);
create_thread(audio_thread, audio_stack, sizeof(audio_stack),
audio_thread_name IF_PRIO(, PRIORITY_BUFFERING)
IF_COP(, CPU, false));
}
void audio_init(void)
{
LOGFQUEUE("audio > audio Q_AUDIO_POSTINIT");
queue_post(&audio_queue, Q_AUDIO_POSTINIT, 0);
}
void voice_init(void)
{
#ifdef PLAYBACK_VOICE
if (!filebuf)
return; /* Audio buffers not yet set up */
if (voice_thread_p)
return;
if (!talk_voice_required())
if (voice_thread_p || !filebuf || voice_codec_loaded ||
!talk_voice_required())
return;
logf("Starting voice codec");
@ -1283,22 +1237,20 @@ static bool voice_request_next_track_callback(void)
static void voice_thread(void)
{
while (1)
{
logf("Loading voice codec");
voice_codec_loaded = true;
mutex_lock(&mutex_codecthread);
set_current_codec(CODEC_IDX_VOICE);
dsp_configure(DSP_RESET, 0);
voice_remaining = 0;
voice_getmore = NULL;
logf("Loading voice codec");
voice_codec_loaded = true;
mutex_lock(&mutex_codecthread);
set_current_codec(CODEC_IDX_VOICE);
dsp_configure(DSP_RESET, 0);
voice_remaining = 0;
voice_getmore = NULL;
codec_load_file(get_codec_filename(AFMT_MPA_L3), &ci_voice);
codec_load_file(get_codec_filename(AFMT_MPA_L3), &ci_voice);
logf("Voice codec finished");
voice_codec_loaded = false;
mutex_unlock(&mutex_codecthread);
}
logf("Voice codec finished");
voice_codec_loaded = false;
mutex_unlock(&mutex_codecthread);
remove_thread(NULL);
} /* voice_thread */
#endif /* PLAYBACK_VOICE */
@ -3465,103 +3417,6 @@ static void audio_reset_buffer(size_t pcmbufsize)
buffer_state = BUFFER_STATE_NORMAL;
}
#ifdef ROCKBOX_HAS_LOGF
static void audio_test_track_changed_event(struct mp3entry *id3)
{
(void)id3;
logf("tce:%s", id3->path);
}
#endif
static void audio_playback_init(void)
{
#ifdef PLAYBACK_VOICE
static bool voicetagtrue = true;
static struct mp3entry id3_voice;
#endif
struct event ev;
logf("playback api init");
pcm_init();
#ifdef AUDIO_HAVE_RECORDING
rec_set_source(AUDIO_SRC_PLAYBACK, SRCF_PLAYBACK);
#endif
#ifdef ROCKBOX_HAS_LOGF
audio_set_track_changed_event(audio_test_track_changed_event);
#endif
/* Initialize codec api. */
ci.read_filebuf = codec_filebuf_callback;
ci.pcmbuf_insert = codec_pcmbuf_insert_callback;
ci.get_codec_memory = codec_get_memory_callback;
ci.request_buffer = codec_request_buffer_callback;
ci.advance_buffer = codec_advance_buffer_callback;
ci.advance_buffer_loc = codec_advance_buffer_loc_callback;
ci.request_next_track = codec_request_next_track_callback;
ci.mp3_get_filepos = codec_mp3_get_filepos_callback;
ci.seek_buffer = codec_seek_buffer_callback;
ci.seek_complete = codec_seek_complete_callback;
ci.set_elapsed = codec_set_elapsed_callback;
ci.set_offset = codec_set_offset_callback;
ci.configure = codec_configure_callback;
ci.discard_codec = codec_discard_codec_callback;
/* Initialize voice codec api. */
#ifdef PLAYBACK_VOICE
memcpy(&ci_voice, &ci, sizeof(struct codec_api));
memset(&id3_voice, 0, sizeof(struct mp3entry));
ci_voice.read_filebuf = voice_filebuf_callback;
ci_voice.pcmbuf_insert = voice_pcmbuf_insert_callback;
ci_voice.get_codec_memory = voice_get_memory_callback;
ci_voice.request_buffer = voice_request_buffer_callback;
ci_voice.advance_buffer = voice_advance_buffer_callback;
ci_voice.advance_buffer_loc = voice_advance_buffer_loc_callback;
ci_voice.request_next_track = voice_request_next_track_callback;
ci_voice.mp3_get_filepos = voice_mp3_get_filepos_callback;
ci_voice.seek_buffer = voice_seek_buffer_callback;
ci_voice.seek_complete = voice_do_nothing;
ci_voice.set_elapsed = voice_set_elapsed_callback;
ci_voice.set_offset = voice_set_offset_callback;
ci_voice.discard_codec = voice_do_nothing;
ci_voice.taginfo_ready = &voicetagtrue;
ci_voice.id3 = &id3_voice;
id3_voice.frequency = 11200;
id3_voice.length = 1000000L;
#endif
codec_thread_p = create_thread(
codec_thread, codec_stack, sizeof(codec_stack),
codec_thread_name IF_PRIO(, PRIORITY_PLAYBACK)
IF_COP(, COP, true));
while (1)
{
queue_wait(&audio_queue, &ev);
if (ev.id == Q_AUDIO_POSTINIT)
break ;
#ifndef SIMULATOR
if (ev.id == SYS_USB_CONNECTED)
{
logf("USB: Audio preinit");
usb_acknowledge(SYS_USB_CONNECTED_ACK);
usb_wait_for_disconnect(&audio_queue);
}
#endif
}
/* initialize the buffer */
filebuf = audiobuf; /* must be non-NULL for audio_set_crossfade */
buffer_state = BUFFER_STATE_TRASHED; /* force it */
audio_set_crossfade(global_settings.crossfade);
audio_is_initialized = true;
sound_settings_apply();
}
#if MEM > 8
/* we dont want this rebuffering on targets with little ram
because the disk may never spin down */
@ -3575,8 +3430,6 @@ static bool ata_fillbuffer_callback(void)
static void audio_thread(void)
{
struct event ev;
/* At first initialize audio system in background. */
audio_playback_init();
while (1)
{
@ -3707,3 +3560,106 @@ static void audio_thread(void)
} /* end while */
}
#ifdef ROCKBOX_HAS_LOGF
static void audio_test_track_changed_event(struct mp3entry *id3)
{
(void)id3;
logf("tce:%s", id3->path);
}
#endif
/* Initialize the audio system - called from init() in main.c.
* Last function because of all the references to internal symbols
*/
void audio_init(void)
{
#ifdef PLAYBACK_VOICE
static bool voicetagtrue = true;
static struct mp3entry id3_voice;
#endif
logf("audio: %s", audio_is_initialized ?
"initializing" : "already initialized");
/* Can never do this twice */
if (audio_is_initialized)
return;
/* Initialize queues before giving control elsewhere in case it likes
to send messages. Thread creation will be delayed however so nothing
starts running until ready if something yields such as talk_init. */
#ifdef PLAYBACK_VOICE
mutex_init(&mutex_codecthread);
#endif
queue_init(&audio_queue, true);
queue_enable_queue_send(&audio_queue, &audio_queue_sender_list);
queue_init(&codec_queue, true);
pcm_init();
#ifdef ROCKBOX_HAS_LOGF
audio_set_track_changed_event(audio_test_track_changed_event);
#endif
/* Initialize codec api. */
ci.read_filebuf = codec_filebuf_callback;
ci.pcmbuf_insert = codec_pcmbuf_insert_callback;
ci.get_codec_memory = codec_get_memory_callback;
ci.request_buffer = codec_request_buffer_callback;
ci.advance_buffer = codec_advance_buffer_callback;
ci.advance_buffer_loc = codec_advance_buffer_loc_callback;
ci.request_next_track = codec_request_next_track_callback;
ci.mp3_get_filepos = codec_mp3_get_filepos_callback;
ci.seek_buffer = codec_seek_buffer_callback;
ci.seek_complete = codec_seek_complete_callback;
ci.set_elapsed = codec_set_elapsed_callback;
ci.set_offset = codec_set_offset_callback;
ci.configure = codec_configure_callback;
ci.discard_codec = codec_discard_codec_callback;
/* Initialize voice codec api. */
#ifdef PLAYBACK_VOICE
memcpy(&ci_voice, &ci, sizeof(ci_voice));
memset(&id3_voice, 0, sizeof(id3_voice));
ci_voice.read_filebuf = voice_filebuf_callback;
ci_voice.pcmbuf_insert = voice_pcmbuf_insert_callback;
ci_voice.get_codec_memory = voice_get_memory_callback;
ci_voice.request_buffer = voice_request_buffer_callback;
ci_voice.advance_buffer = voice_advance_buffer_callback;
ci_voice.advance_buffer_loc = voice_advance_buffer_loc_callback;
ci_voice.request_next_track = voice_request_next_track_callback;
ci_voice.mp3_get_filepos = voice_mp3_get_filepos_callback;
ci_voice.seek_buffer = voice_seek_buffer_callback;
ci_voice.seek_complete = voice_do_nothing;
ci_voice.set_elapsed = voice_set_elapsed_callback;
ci_voice.set_offset = voice_set_offset_callback;
ci_voice.discard_codec = voice_do_nothing;
ci_voice.taginfo_ready = &voicetagtrue;
ci_voice.id3 = &id3_voice;
id3_voice.frequency = 11200;
id3_voice.length = 1000000L;
#endif
/* initialize the buffer */
filebuf = audiobuf; /* must be non-NULL for audio_set_crossfade */
/* audio_reset_buffer must to know the size of voice buffer so init
voice first */
talk_init();
codec_thread_p = create_thread(
codec_thread, codec_stack, sizeof(codec_stack),
codec_thread_name IF_PRIO(, PRIORITY_PLAYBACK)
IF_COP(, COP, true));
create_thread(audio_thread, audio_stack, sizeof(audio_stack),
audio_thread_name IF_PRIO(, PRIORITY_BUFFERING)
IF_COP(, CPU, false));
audio_set_crossfade(global_settings.crossfade);
audio_is_initialized = true;
sound_settings_apply();
audio_set_buffer_margin(global_settings.buffer_margin);
} /* audio_init */

6
apps/talk.c
View file

@ -182,12 +182,6 @@ static void load_voicefile(void)
load_size = file_size;
#endif
#ifndef SIMULATOR
/* Wait until the audio is initialized before continuing */
while(!audio_is_initialized)
sleep(HZ/100);
#endif
got_size = read(filehandle, audiobuf, load_size);
if (got_size != load_size /* failure */)
goto load_err;

2
firmware/SOURCES
View file

@ -161,7 +161,9 @@ tuner_philips.c
#endif /* CONFIG_TUNER */
/* Sound */
#if CONFIG_CODEC != SWCODEC
mp3_playback.c
#endif /* CONFIG_CODEC != SWCODEC */
mp3data.c
sound.c