On Fri, Oct 24, 2014 at 9:10 AM, Fabio Estevam festevam@gmail.com wrote:
On Fri, Oct 24, 2014 at 12:04 PM, Mark Brown broonie@kernel.org wrote:
On Fri, Oct 24, 2014 at 10:20:29AM -0200, Fabio Estevam wrote:
On Fri, Oct 24, 2014 at 2:42 AM, Jonathan Bennett <
jbscience87@gmail.com> wrote:
- ssi_clk = devm_clk_get(&codec_dev->dev, "cko");
- if (IS_ERR(ssi_clk)) {
- pr_err("Could not get "cko" clock \n");
- return PTR_ERR(ssi_clk);
I am not sure that adding the clock handling here is the correct
approach.
The dts you use looks like:
codec: wm8731@1a { compatible = "wlf,wm8731"; reg = <0x1a>; clocks = <&clks 173>, <&clks 158>, <&clks 201>, <&clks 200>; clock-names = "pll4", "imx-ssi.1", "cko", "cko2";
,which does not match the current bindings for the wm8731. The wm8731 driver could be extended to handle the clocks if needed, instead of handling them via machine driver.
The WM8731 doesn't have anything like that number of clock inputs so hanging them off the WM8731 seems very confused.
Yes, exactly. If we need to pass a clock to WM8731 it should be MCLK only.
The pll4, ssi clocks are internal mx6 clocks, so the proposed clock handling should be reworked.
All 4 clocks were in that dts as a result of tinkering with both slave and master mode. We are actually only using the first two. That said, I think you're right. I think the ones we use should be moved to the sound {} block of the dts. I'll get these changes fixed up and send in the next version.
Oh, also, this patch is a bit big for the mailing list, what is the preferred method for sending in large patches to look at? I can attach, throw it on github, put it on an http server, etc.