Hi Mark,
On 03/19/2014 06:28 PM, Mark Brown wrote:
On Wed, Mar 19, 2014 at 01:57:23PM +0800, Bo Shen wrote:
On 03/17/2014 07:55 PM, Mark Brown wrote:
If this is a clock for the CODEC it should be documented as part of the
binding for the CODEC and connected to the CODEC in the device tree
rather than being part of a machine driver binding.
This is a optional clock for CODEC which depends on hardware design. There
are 3 options for this clock, wm8904 as an example.
1. Using internal FLL, so won't use this clock.
2. Using external oscillator, no need to retrieve this clock.
3. Using SoC provide this clock (we use this case).
After considering these 3 options, if we put this into CODEC driver to do
it, I think it will be more complicate to do logic judgement. Do you think
so?
There shouldn't be any meaningful complexity from the above cases -
cases 2 and 3 are the same and if the clock isn't used at all then it
can be omitted. If the FLL is clocked from MCLK then the CODEC driver
should be able to work out how to configure it easily, the device isn't
like a digital hub CODEC with lots of clocking options.
For this, in my mind, I think we need add following parameters in DT.
1. sysclk_src_from_fll --> we need do nothing.
2. sysclk_src_from_mclk
2.1 use_external_xtal --> we need clock_frequency
2.2 !use_external_xtal --> we need retrieve clock and clock_frequency.
So, the dt may looks like:
for case 1:
wm8904: wm8904@1a {
reg = <0x1a>;
sysclk_src_from_fll;
}
for case 2.1:
wm8904: wm8904@1a {
reg = <0x1a>;
sysclk_src_from_mclk;
use_external_xtal;
clock_frequency = 12000000;
}
for case 2.2:
wm8904: wm8904@1a {
reg = <0x1a>;
sysclk_src_from_mclk;
clocks = <&pck0>;
clock-names = "mclk";
clock_frequency = 32768;
}
Does this acceptable? Or any other better suggestion for this?
Best Regards,
Bo Shen
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html