On Fri, Apr 17, 2009 at 03:55:07PM +0300, Peter Ujfalusi wrote:
Note that I'm not using the just introduced symmetric_rates feature of ASoC introduced by Mark Brown recently, since I think that also suffers from this problem (or not, I have to try it out first to be sure about it)
Yes, pretty much all the code to handle this sort of issue has that problem in some form. There's issues with the constraints not getting seen in time - people had a look at it when we started doing constraints like this properly.
The symmetric rate setting tries to cover it a bit by enforcing the last constraint that was seen so you only get the zero problem on first open. For a lot of applications this actually ends up doing the right thing, though obviously it's not ideal.