On Sun, May 24, 2009 at 11:28:15AM -0400, Jon Smirl wrote:
On Sun, May 24, 2009 at 7:11 AM, Mark Brown
Why do you wish to do this - ASoC does not require or use DAI registration for AC97 CODECs?
Then what is wrong with my binding code? If I take out the registration my bind fails.
It appears that the problem here is that your CPU DAI isn't marked as an ac97 DAI by having ac97_control set so the core expects the codec to instantiate prior to the card. Setting ac97_control ought to fix the problem, but obviously I can't test.
When looking at problems like this it's worth taking a step back and looking at why the existing code is the way that it is when considering if you've found the right fix. In this case AC97 is fairly widely used and there are also a number of WM9712 users but neither WM9712 or any of the other AC97 CODEC drivers ever register their DAIs. That should be an indication that either things probably should work without a change to existing code or things are completely broken and something wider than a change in a single driver is required in order to get things working.