ao_alsa: handle XRUNs separately from other errors

According to ALSA doxy, EPIPE is a synonym to SND_PCM_STATE_XRUN,
and that is a state that we should attempt to automatically recover
from. In case recovery fails, log an error and return zero.

A warning message will still be output for each XRUN since those
are not something we should generally be receiving.
This commit is contained in:
Jan Ekström
2018-09-27 20:39:49 +03:00
committed by sfan5
parent 3218a58082
commit fdc952486a

View File

@@ -947,10 +947,15 @@ static int get_space(struct ao *ao)
snd_pcm_sframes_t space = snd_pcm_avail(p->alsa);
if (space < 0) {
if (space == -EPIPE) {
MP_WARN(ao, "ALSA XRUN hit, attempting to recover...\n");
int err = snd_pcm_prepare(p->alsa);
CHECK_ALSA_ERROR("Unable to recover from under/overrun!");
return p->buffersize;
}
MP_ERR(ao, "Error received from snd_pcm_avail (%ld, %s)!\n",
space, snd_strerror(space));
if (space == -EPIPE) // EOF
return p->buffersize;
// request a reload of the AO if device is not present,
// then error out.