[alsa-devel] [PATCH v5 1/2] ALSA: hda: add Intel DSP configuration / probe code

Pierre-Louis Bossart pierre-louis.bossart at linux.intel.com
Fri Oct 25 16:10:32 CEST 2019


>> Jaroslav, I found a case where the logic doesn't seem to work as 
>> intended.
>>
>> the code currently scans through the table looking for a PCI ID match,
>> and a DMI match for quirks.
>>
>> On a KBL NUC device with the DSP enabled, we go through the entire list
>> without finding a match, which means we return the _ANY flag. so if we
>> try to probe the snd-soc-skl driver, the checks will allow the driver to
>> continue, despite this configuration not being supported.
>>
>> It was my understanding that we would always load the legacy driver
>> *unless* there is a match in the table? If yes, there we can change the
>> default this way:
>>
>> @@ -333,7 +355,7 @@ int snd_intel_dsp_driver_probe(struct pci_dev *pci)
>>           /* find the configuration for the specific device */
>>           cfg = snd_intel_dsp_find_config(pci, config_table,
>> ARRAY_SIZE(config_table));
>>           if (!cfg)
>> -               return SND_INTEL_DSP_DRIVER_ANY;
>> +               return SND_INTEL_DSP_DRIVER_LEGACY;
>>
>> Or we need to add a case where we force the legacy in the table for each
>> platform.
> 
> I would prefer to add another entry to the configuration table without 
> .flags (or .flags = 0) with a comment, why the legacy driver is selected 
> by default even if the DSP is present.

Ah yes, flags == 0 would mean legacy indeed. Will send a patch, thanks 
for the idea.


More information about the Alsa-devel mailing list