Re: [PATCH 0/4] x86: Trust firmware a bit more about presence of 8042

From: Takashi Iwai
Date: Mon Dec 12 2016 - 09:30:37 EST

On Fri, 09 Dec 2016 21:57:37 +0100,
Dmitry Torokhov wrote:
> Hi,
> Historically we did not trust PNP data regarding keyboard controllers on
> X86, but more and more boards get upset with us if they try to tell us that
> there is no keyboard controller and we still go and try to poke at where we
> think it might be. To work around this issue let's have a bit more faith in
> BIOS data, and if [lack] of PNP devices for mouse and keyboard matches whet
> firmware (basically ACPI FADT) tells us, let's abort i8042 probe.
> We add a new flag (enum) to x86_platform.legacy structure so we can
> distinguish between cases where platform/subarch never has 8042 (such as
> MID platform) and cases where firmware says that it is not there, so that
> i8042 driver can either abort immediately or go and check for presence of
> PNP devices. We also remove x86_platform.i8042_detect() as it is no longer
> used (platforms can set value of x86_platform.legacy.i8042 as needed in
> quirks).
> If you are OK with arch/x86 changes please apply together with the input
> part.

I tested and worked fine. For all patches:
Tested-by: Takashi Iwai <tiwai@xxxxxxx>