A few months ago we'd digged jackd code in #alsa IRC trying to find why it fails with "ALSA: could not start playback (File descriptor in bad state)" when run on top "type bluetooth" pcm. The reason was in a single line (both in jack1/drivers/alsa/alsa_driver.c and jack2/linux/alsa/alsa_driver.c): if ((err = snd_pcm_start (driver->playback_handle)) < 0) { which always fails after snd_pcm_mmap_commit() filling the buffer, because pcm's already in RUNNING state and can't be started again: > http://www.alsa-project.org/alsa-doc/alsa-lib/pcm.html > For playback, if samples in ring buffer is equal or greater than > the start threshold parameters and the stream is not running, > the stream will be started automatically The following small patch fixes it for both jack1 and jack2: --- alsa_driver.c +++ alsa_driver.c @@ -1055,7 +1055,8 @@ alsa_driver_start (alsa_driver_t *driver driver->user_nperiods * driver->frames_per_cycle); - if ((err = snd_pcm_start (driver->playback_handle)) < 0) { + if (snd_pcm_state(driver->playback_handle) != SND_PCM_STATE_RUNNING + && (err = snd_pcm_start (driver->playback_handle)) < 0) { jack_error ("ALSA: could not start playback (%s)", snd_strerror (err)); return -1;