Device is not in proper state - after RC1

[yoda]

RC1 does not have this problem - it starts from RC2

From logs for almost all hardware I have seen, from RC2 and on we always seem to have left device in an non proper state. a short example from alsahlp$:

Hardware parameters: sample rate 22050, data type 2, channels 1, period size 1024, periods 32 Device is not in proper state: 1. Calling prepare

All devices seems to play ok - but in RC1 it seems like "prepare" was never called, so something must have been changed, since we now always are in non proper state

[psmedley]

Changeset 345 fixes this.