On Sun, Sep 6, 2009 at 10:22 AM, Mark Brownbroonie@opensource.wolfsonmicro.com wrote:
On Sun, Sep 06, 2009 at 09:09:41AM +0900, jassi brar wrote:
- case WM8580_MCLKRATIO:
- case WM8580_BCLKRATIO:
There should be no need for machine drivers to configure any of this, there's already enough information in the CODEC driver to figure out this configuration without their help.
Not all combinations of BFS and RFS are possible at all LRCLK values for CPU_DAI, not for Samsung SoCs those I am aware of. So, i think the machine driver, as a link between Codec and Cpu Dai, has the job of figuring out values for BFS and RFS that are compatible to _both_ the CPU and the CODEC. If CODEC driver, with disregard to the CPU, chooses values for BCLK/MCLK-Ratio that are incompatible with CPU, i don't know how that will work out. Please correct, if i am missing something.
As mentioned previously this driver could use a bit of modernisation in general, including with regard to clocking, which I had already been planning to do as part of working on the SMDK boards. As a result of some work on another chip with a very similar clocking structure I have a rough draft of what needs doing for the clocks already. This is part of why in my earlier mail I had recommended focusing on the CPU DAI drivers for now - the lack of them is the biggest obstacle to getting that work cleaned up and tested.
I get your recommendation, but i am already in the middle of submitting I2S machine driver for 6410-WM8580 mainline and that is one of the tasks i am assigned. Ofcourse, after 6410-wm8580 driver is all up and available to our users, i can switch to AC97.