Move the workaround for the VT8251 up a bit, and check for STAT_EOL
rather than STAT_ACTIVE. This resolves issues some people were having
with certain ALSA clients (and allows the STAT_ACTIVE check to do what
it was intended to do).
This change was suggested by Andrew Daviel.
Signed-off-by: Bastiaan Jacques <b.jacques@planet.nl>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
if (!status)
status = inb(VIADEV_REG(viadev, OFFSET_STATUS));
+ /* An apparent bug in the 8251 is worked around by sending a
+ * REG_CTRL_START. */
+ if (chip->revision == VIA_REV_8251 && (status & VIA_REG_STAT_EOL))
+ snd_via82xx_pcm_trigger(substream, SNDRV_PCM_TRIGGER_START);
+
if (!(status & VIA_REG_STAT_ACTIVE)) {
- /* An apparent bug in the 8251 is worked around by sending
- * a REG_CTRL_START. */
- if (chip->revision == VIA_REV_8251)
- snd_via82xx_pcm_trigger(substream,
- SNDRV_PCM_TRIGGER_START);
- else {
- res = 0;
- goto unlock;
- }
+ res = 0;
+ goto unlock;
}
if (count & 0xffffff) {
idx = count >> 24;