[alsa-devel] [PATCH 7/7] ASoC: Intel: common: add quirk for APL RVP boards

[Pierre-Louis Bossart]

On 11/1/18 9:11 AM, Andy Shevchenko wrote:
> On Wed, Oct 31, 2018 at 08:07:18PM -0500, Pierre-Louis Bossart wrote:
>> For some reason the RVP/LeafHill SSDT exposes an INT34C3 ID which is
>> used on other boards to point to the TDF8532 amplifier. Yay BIOS.
>>
>> Add a DMI-quirk to ignore this ID and check for other valid machine
>> driver descriptors.
>> +static unsigned long apl_machine_id;
>> +
>> +#define APL_RVP  1
>> +
>> +static int apl_rvp_quirk_cb(const struct dmi_system_id *id)
>> +{
>> +	apl_machine_id = APL_RVP;
>> +	return 1;
>> +}
>> +
>> +static const struct dmi_system_id apl_table[] = {
>> +	{
>> +		.callback = apl_rvp_quirk_cb,
>> +		.matches = {
>> +			DMI_MATCH(DMI_SYS_VENDOR, "Intel Corp."),
>> +			DMI_MATCH(DMI_BOARD_NAME, "Apollolake RVP1A"),
>> +		},
>> +	},
>> +	{},
> Terminator entry doesn't need a comma. It would prevent from (unlikely)
> mistakes when extending the list.
Indeed. I don't know how many tables use this though... will update.
>
>> +};
>> +
>> +static struct snd_soc_acpi_mach *apl_quirk(void *arg)
>> +{
>> +	struct snd_soc_acpi_mach *mach = arg;
>> +
>> +	dmi_check_system(apl_table);
> Can't we just use driver_data of the table above and get rid of at least global variable?

Yes. The code is pretty much copy-pasted from other BYT quirks but can 
be written in a better way

Will send a v2 later today, thanks Andy for the suggestion.


>
>> +
>> +	if (apl_machine_id == APL_RVP)
>> +		return NULL;
>> +	else
>> +		return mach;
>> +}