On Wed, 2008-11-26 at 21:16 +0000, Mark Brown wrote:
On Wed, Nov 26, 2008 at 12:33:12PM -0800, David Brownell wrote:
And maybe more; that's just me summarizing unconnected pins, not the datasheet. There are also cost-reduced parts with less audio capability.
...
I make no claim about audio expertise -- the nearest I come to it is observing a few years back that there was a huge framework hole, which ASoC seems to fill -- but I wonder if it shouldn't suffice just to provide a mask of capabilities that are wired up on a given board.
It's possible - if the differences can be handled by marking some pins as connected or disconnected and possible specifying a different system clock rate to the chip then that approach was what I was expecting. If there is more substantial differences such as having the clocking arranged differently then separate drivers start to make more sense.
I had considered the differences among every boards. If you want add some board specific dapm widgets or other features, you can use the extended interface to add them. If a board has too much differences, you can separate it from this driver and implement it independently.
Now there are four same driver files in alsa-devel tree and there are another three same drives waiting to be pushed. So I think this shared diver is needed.
After reading all of the suggestions, I reworked the driver and will send it again. This new version has a better difference-tolerance ability.
Stanley.