[alsa-devel] [PATCH 2/2] ASoC: Intel: add machine driver for BYT/CHT + ES8316

[Pierre-Louis Bossart]

On 5/15/17 1:07 PM, Daniel Drake wrote:
> On Wed, May 10, 2017 at 12:49 PM, Pierre-Louis Bossart
> <pierre-louis.bossart at linux.intel.com> wrote:
>> You may want to add a comment that this patch is based on the use of SSP2
>> and hence cannot work on BYT-CR devices.
>
> I'm not quite clear on the terminology here. What does BYT-CR mean exactly?

a cost-reduced version where the packaging is different with only SSP0 
and SSP2 exposed.

>
> I named the driver with "byt" in addition to "cht" as I was hoping it
> can support the baytrail devices mentioned here:
> https://bugzilla.kernel.org/show_bug.cgi?id=189261#c71
>
> I also see that bytcht_da7213 is documented as supporting Baytrail
> devices and only uses SSP2. But bytcr_rt5640 does switch to SSP0 on
> valleyview, and I could do the same.
>
> Not sure which way I should go here at least without more info from
> the reporter, let me know if you have any suggestions.

just add a note that SSP0 routing is not supported at the moment, no 
reason to add code if you can't test it.

>
>>> +static const struct snd_kcontrol_new byt_cht_es8316_controls[] = {
>>> +       SOC_DAPM_PIN_SWITCH("Headphone"),
>>> +       SOC_DAPM_PIN_SWITCH("Microphone 1"),
>>> +       SOC_DAPM_PIN_SWITCH("Microphone 2"),
>>
>>
>> do you really need to control the two analog mics independently? And along
>> the same lines, should there be a path for DMICS?
>
> The ES8316 datasheet is clear that it supports 2 independent analog
> stereo mics, and the codec driver exposes the relevant controls for
> that, so I added two here. There is a mux so you have to choose which
> one you feed into the ADC, but (if you wanted to) you could e.g. feed
> the other into the HP output mixer, and use both simultaneously.
>
> The product I'm working with only uses one of them though, so I could
> delete the other one if you think that makes sense?

I am not sure what the recommendation is. I've always seen a single 
'mic' or 'int mic' switch being used regardless of the number of mics. 
Usually you have additional routing controls to define which mics are 
used, this control seems to be a higher level kill switch of sorts.

>
> The datasheet does say "Support digital mic" but it makes almost no
> other mention of DMIC support, not even which of the 4 mic input pins
> you would use for that. The codec driver does expose a single
> SND_SOC_DAPM_INPUT("DMIC") though so I could very easily add a route
> here, although I would have no way of testing it.

ok, fair enough. adding a note of what was tested is helpful in general.

>
> Thanks for your help
>
> Daniel
> _______________________________________________
> Alsa-devel mailing list
> Alsa-devel at alsa-project.org
> http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
>