We still see the occasional timeout waiting for busy to clear. As the spec is contradictory, and we know that the current implementation doesn't work, try an alternative interpretation from the spec. This one appears to work - I have yet to find any issue with it during my testing over several months.
Signed-off-by: Russell King rmk+kernel@arm.linux.org.uk --- sound/soc/kirkwood/kirkwood-i2s.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/sound/soc/kirkwood/kirkwood-i2s.c b/sound/soc/kirkwood/kirkwood-i2s.c index b601ad680d7b..e98650c01eba 100644 --- a/sound/soc/kirkwood/kirkwood-i2s.c +++ b/sound/soc/kirkwood/kirkwood-i2s.c @@ -238,7 +238,7 @@ static int kirkwood_i2s_play_trigger(struct snd_pcm_substream *substream, uint32_t ctl, value;
ctl = readl(priv->io + KIRKWOOD_PLAYCTL); - if (ctl & KIRKWOOD_PLAYCTL_PAUSE) { + if ((ctl & KIRKWOOD_PLAYCTL_ENABLE_MASK) == 0) { unsigned timeout = 5000; /* * The Armada510 spec says that if we enter pause mode, the