Re: [PATCH v2 0/5] ASoC: Intel: ACPI support and machines

[Jarkko Nikula <jarkko.nikula@linux.intel.com>]

Hi

On 11/04/2014 10:25 AM, Vinod Koul wrote:
> On Tue, Nov 04, 2014 at 10:09:40AM +0200, Jarkko Nikula wrote:
>>
>> I think there is need to have some consolidation with current
>> byt-rt5640 and byt-max98090. Since existing drivers and this set are
>> using the same ACPI IDs "80860F28" for the ADSP and "10EC5640" codec
>> ACPIID for the machine driver probing these drivers won't work
>> together.
> Liam's suggestion was to remove the current driver when all the bits are
> upstream

Yeah, I'm not against it but we must avoid regressions and perhaps try 
to do seamless handover if possible. It's always more nice to find 
regression in a small than big patch when bisecting :-)

>> One difference at the moment is that byt-rt5640 machines out there
>> have differences in audio routings but those are not too difficult
>> to handle in the same machine driver.
>>
>> Bigger question how to handle platform part? I.e. can these use the
>> same drivers and ADSP firmware? If they can, is the firmware
>> available for distributions to take?
> The two firmware are *not* compatible and that is the big reason why we
> can't reuse the parts available

Ok, so firmware appears to works on other than CR Baytrails too. Is it 
available in linux-firmware since if it's not when changing the platform 
drivers there will be obvious regressions to users?

>> Optimally all machines out there would use the same drivers and get
>> support for compressed audio but meanwhile we should not break
>> existing machines.
> Yes, I have tested this one on T100 and doesn't seem to break anything.
> Please do test and revert if you find issues on any other available board (I
> don't have any other board)
>
Well, those audio routing differences require to have DMI quirks. E.g. 
T100 has analogue internal mic, Dell Venue 8 Pro 5830 has digital mic 
connected to DMIC2 of the RT564x and some have it connected to DMIC1.

-- 
Jarkko