Fixed alsa dev read pcm bug #4206
Conversation
|
|
||
| snd_pcm_uframes_t avail; | ||
| avail = snd_pcm_avail(pcm); | ||
| if (avail > 0 && avail <= nframes) { |
There was a problem hiding this comment.
shouldn't this be >=0 and < nframes? We shouldn't need to sleep and continue when we already have nframes.
There was a problem hiding this comment.
when snd_pcm_prepare called,the snd_pcm_state is SND_PCM_STATE_PREPARED,snd_pcm_avail will return 0 always,need call snd_pcm_readi or snd_pcm_stop to change state to SND_PCM_STATE_RUNNING. i will update code with snd_pcm_state check later.
|
I tried to reproduce it here using a USB/removable sound card during a call/sound card is used, however it doesn't block instead printing this on the log: Any specific scenario to trigger this issue? |
|
set playback device plughw:usb sound card,and set capture device default. |
freeforf
force-pushed
the
fixbugs/alsa_dev
branch
from
ff98f3c to
6c329bb
Compare
Still couldn't reproduce it. Based on the doc the |
|
i insert two same usb soundcard,if it doesn't work too,maybe my soundcard bugs. |
|
Perhaps you can consider submitting your issue to https://github.com/alsa-project/alsa-lib/issues (https://www.alsa-project.org/wiki/Bug_Tracking). |
remove sound card device when capture audio data, the alsa audio cache may less than nframes, snd_pcm_readi wait for audio data forever.so when close the audio device ,the ca_thread_func can`t quite.use snd_pcm_avail to make sure audio cache has enough audio data.