Dne 08. 12. 20 v 19:23 Mukunda,Vijendar napsal(a):
On 08/12/20 11:51 pm, Mukunda,Vijendar wrote:
On 08/12/20 11:27 pm, Takashi Iwai wrote:
On Tue, 08 Dec 2020 19:06:21 +0100, Mukunda,Vijendar wrote:
On 08/12/20 11:10 pm, Mark Brown wrote:
On Tue, Dec 08, 2020 at 05:24:32PM +0100, Takashi Iwai wrote:
BTW, both Raven and Reonir drivers point to the very same PCI ID, and both drivers will be probed for this machine (and both to be skipped).
Ugh, that's not good. It's not even super obvious from the code that this is happening. Seems like it should be one core driver which instantiates the components for Raven and Reonir as appropriate, the PCI driver is pretty thin at present anyway.
Raven and Renoir has same PCI ID but both platforms have different revision ID. Raven platform revision id is 0x00 where as for Renoir it is 0x01.
But your drivers don't check the revision ID, as far as I see?
The linux PCI driver doesn't distinguish the revision id at the matching time, unfortunately.
Takashi
Apart from Revision ID difference, There are few hardware differences specific to ACP IP. ACP IP hardware versions are different for Raven and Renoir. Unfortunately we don't have specific logic to distinguish ACP hardware versions for Raven and Renoir.
But build wise both Raven and Renoir uses different Kconfig options.
We need to build all drivers for the universal distros to one kernel. The Kconfig does not help here.
Jaroslav