Alan Horstmann wrote:
On Tuesday 30 September 2008 15:42, Frank Mandarino wrote:
See attached.
I am a bit unclear about the implications of this, when it was first discussed and now there is a definite patch to remove eti_b1_wm8731.c
We based on this file to create an AT91<->wm8731 system; how is it proposed that the codec is glued to the board in the future? Or is most of the code superfluous anyway?
Alan
The eti_b1_wm8731.c is the board-specific ASoC machine driver for our proprietary ETI-B1 board. It was only added to the kernel as an example of an AT91 machine driver for others to use when developing a machine driver for their own boards.
ASoC requires you to create your own machine driver to glue the codec to the board. The code is not superfluous.
If you have developed an AT91<->wm8731 board which you are planning to sell, it makes sense to add your machine driver to the kernel so that others can build a kernel for your board. If you are not selling your board, then it probably shouldn't be in the kernel and you should maintain your machine driver privately.
This real issue is that only our company can test ASoC changes against our board, and we don't have always have the resources to do this in a timely fashion. So I would rather not require others who make improvements to ASoC to have to make changes to eti_b1_wm8731.c when they are unable to test them.
Regards, ../fam