Re: [PATCH 5.4 regression fix] Input: soc_button_array - partial revert of support for newer surface devices

From: Maximilian Luz
Date: Sat Oct 05 2019 - 11:01:38 EST


Hi, again

On 10/5/19 3:20 PM, Hans de Goede wrote:
Ok, on x86 the GPIO drivers really should all be builtin because
various ACPI methods including device D0 / D3 (power-on/off) methods
may depend on them. So normally this should never happen.

If this (-EPROBE_DEFER on surface devices) somehow still is happening
please let me know and we will figure something out.

I have never personally experienced this, only received reports which
indicated this and that the change (as well as manually reloading
soc_button_array) fixed it. I will come back to you if I hear anything
in regards to this again.

I have now also tested your patch on the Surface Book 2. Does not cause
any issues as far as I can tell.

Tested-by: Maximilian Luz <luzmaximilian@xxxxxxxxx>

And if that is needed/wanted

Acked-by: Maximilian Luz <luzmaximilian@xxxxxxxxx>

Regards,

Maximilian