| Message ID | 20170613141812.21220-5-tiwai@suse.de (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Jun 13 2017 23:18, Takashi Iwai wrote: > Calling PREPARE ioctl to the stream in either PAUSED or SUSPENDED > state may confuse some drivers that don't handle the state properly. > Instead of fixing each driver, PCM core should take care of the proper > state change before actually trying to (re-)prepare the stream. > Namely, when the stream is in PAUSED state, it triggers PAUSH_RELEASE, s/PAUSH_RELEASE/PAUSE_RELEASE/ > and when in SUSPENDED state, it triggers STOP, before calling prepare > callbacks. > > Signed-off-by: Takashi Iwai <tiwai@suse.de> > --- > sound/core/pcm_native.c | 11 +++++++++++ > 1 file changed, 11 insertions(+) > > diff --git a/sound/core/pcm_native.c b/sound/core/pcm_native.c > index 69cf9b02ac70..0941b9c92b3f 100644 > --- a/sound/core/pcm_native.c > +++ b/sound/core/pcm_native.c > @@ -1684,6 +1684,17 @@ static int snd_pcm_prepare(struct snd_pcm_substream *substream, > else > f_flags = substream->f_flags; > > + snd_pcm_stream_lock_irq(substream); > + switch (substream->runtime->status->state) { > + case SNDRV_PCM_STATE_PAUSED: > + snd_pcm_pause(substream, 0); > + /* fallthru */ > + case SNDRV_PCM_STATE_SUSPENDED: > + snd_pcm_stop(substream, SNDRV_PCM_STATE_SETUP); > + break; > + } > + snd_pcm_stream_unlock_irq(substream); > + > return snd_pcm_action_nonatomic(&snd_pcm_action_prepare, > substream, f_flags); > } Regards Takashi Sakamoto
diff --git a/sound/core/pcm_native.c b/sound/core/pcm_native.c index 69cf9b02ac70..0941b9c92b3f 100644 --- a/sound/core/pcm_native.c +++ b/sound/core/pcm_native.c @@ -1684,6 +1684,17 @@ static int snd_pcm_prepare(struct snd_pcm_substream *substream, else f_flags = substream->f_flags; + snd_pcm_stream_lock_irq(substream); + switch (substream->runtime->status->state) { + case SNDRV_PCM_STATE_PAUSED: + snd_pcm_pause(substream, 0); + /* fallthru */ + case SNDRV_PCM_STATE_SUSPENDED: + snd_pcm_stop(substream, SNDRV_PCM_STATE_SETUP); + break; + } + snd_pcm_stream_unlock_irq(substream); + return snd_pcm_action_nonatomic(&snd_pcm_action_prepare, substream, f_flags); }
Calling PREPARE ioctl to the stream in either PAUSED or SUSPENDED state may confuse some drivers that don't handle the state properly. Instead of fixing each driver, PCM core should take care of the proper state change before actually trying to (re-)prepare the stream. Namely, when the stream is in PAUSED state, it triggers PAUSH_RELEASE, and when in SUSPENDED state, it triggers STOP, before calling prepare callbacks. Signed-off-by: Takashi Iwai <tiwai@suse.de> --- sound/core/pcm_native.c | 11 +++++++++++ 1 file changed, 11 insertions(+)