Hey, On Wed, 13 Jan 2021, Pierre-Louis Bossart wrote: > On 1/13/21 10:20 AM, Takashi Iwai wrote: > > > +++ b/sound/soc/sof/sof-pci-dev.c > > > @@ -512,6 +512,8 @@ static const struct pci_device_id sof_pci_ids[] = { > > > #if IS_ENABLED(CONFIG_SND_SOC_SOF_ALDERLAKE) > > > { PCI_DEVICE(0x8086, 0x7ad0), > > > .driver_data = (unsigned long)&adls_desc}, > > > + { PCI_DEVICE(0x8086, 0x51c8), > > > + .driver_data = (unsigned long)&tgl_desc}, > > > > Is it correct? It's referring to TGL thing and it's not enabled with > > CONFIG_SND_SOC_SOF_ALDERLAKE. > > Yes that should be CONFIG_SND_SOC_SOF_TIGERLAKE. ack Takashi, 'tgl_desc' is correct but the #if guards are not correct and will result in build fail if you disable just CONFIG_SND_SOC_SOF_TIGERLAKE. In SOF CI we of course have all platforms enabled (or none), so this is not caught. I'll send a V2 tomorrow (before Pierre completes the bigger rework to sof-pci-dev). > I am starting to think as part of the Kconfig rework we might get rid of this > granularity. It's not very useful to allow for a single PCI ID to be selected, Ack, my votes for this :) Br, Kai