On Mon, Apr 18, 2016 at 4:46 AM, Mark Brown <broo...@kernel.org> wrote:
> On Mon, Apr 18, 2016 at 06:20:00PM +0800, Xing Zheng wrote:
>> On 2016年04月18日 18:03, Xing Zheng wrote:
>
>> sound {
>> ......
>> simple-audio-card,dai-link@0 {
>> format = "i2s";
>> cpu {
>> sound-dai = <&i2s0>;
>> };
>>
>> codec {
>> sound-dai = <&codec>;
>>
>> simple-audio-card,codec-jack =
>> "JACK_HEADSET",
>> "JACK_BTN_0",
>> "JACK_BTN_1",
>> "JACK_BTN_2",
>> "JACK_BTN_3";
>> };
>> };
>
> This seems like it's only half the job and worryingly close to Linux
> internals. In particular the fact that the binding is specific to
> simple-card and the fact that it's being placed on the CODEC (rather
> than a separate object that the CODEC references) so that we can't
> combine multiple devices are both a concern.
>
> Dylan Reid did have an earlier go at defining a binding for this:
>
> http://thread.gmane.org/gmane.linux.alsa.devel/138906
>
> which went through a couple more iterations but the work on that seemed
> to die off a bit.
Thanks for forwarding this. I did intend to get back to that. But,
almost a year later, obviously I haven't.
Xing if you want to pick this up it would be great. Lars had some
good points on that original discussion, which I think we can
integrate into a single solution.
