2 Mar
2016
2 Mar
'16
4:04 a.m.
On Tue, Feb 23, 2016 at 10:51:14AM +0000, Liam Girdwood wrote:
The only problem we have with using a kcontrol to define params is that we already have an existing kernel ioctl and public C API to define SW/HW params. This would add another mechanism, unless each new kcontrol allows modification of a single param ?
It's to an extent a choice between two evils here - if we've got something that looks like you should be able to play audio through it but you can't actually play audio through it and need to write a custom application to control it then how much are we winning in consistency? On the other hand the mechanisms for discovering and running through constraints that PCMs currently have are useful.