Re: [PATCH 1/3] ASoc: amd: create platform device for VG machine driver

["Mukunda,Vijendar" <vijendar.mukunda@amd.com>]

On 10/11/21 11:59 PM, Mark Brown wrote:
> On Mon, Oct 11, 2021 at 11:48:40AM +0530, Mukunda,Vijendar wrote:
>> On 10/11/21 11:19 PM, Mark Brown wrote:
> 
>>>> +		pdevinfo[3].name = "acp5x_nu8821_cs35l41_mach";
> 
>>> This appears to unconditionally assume that any machine with this
>>> hardware is going to have exactly the same CODEC/amp combination - that
>>> seems like it's going to go badly at some point.  Shouldn't there be
>>> some other check that we're instantiating the right machine driver?
> 
>> we will make the platform device as generic one something like "acp5x_mach".
> 
>> Currently we have only one target platform with above codec combinations
>> for this APU series.
> 
>> When we get new codecs combinations to support, we will address it in
>> machine driver by adding DMI checks for distinguishing codec combinations.
> 
> If that's the case then this machine driver that's being instantiated is
> poorly named, and I expect you're going to get issues with assuming a
> default here and trying to instantiate this machine on unsuitable
> hardware.  It'd be better to at least put a bit of the framework in now
> and positively indentify systems that can run this machine driver.

Will address it by adding DMI checks in machine driver code.
> 
> It really would be good if ACPI system vendors were to adopt a more
> standards based approach to platform enumeration, this really isn't
> good.  Something more standards based like DT has would be much more
> scalable.
>