Re: [REGRESSION][BISECTED] Dell Precision 7780 wakes up on its own from suspend

[Mario Limonciello <superm1@kernel.org>]

On 7/18/2025 10:13 AM, Werner Sembach wrote:
> Hi,
> 
> Am 18.07.25 um 01:29 schrieb Askar Safin:
>>   ---- On Mon, 30 Jun 2025 22:40:28 +0400  Mario 
>> Limonciello<superm1@kernel.org> wrote ---
>>   > Looks like your interrupt 14 is ACPI device INTC1085:00.
>>   >
>>   > Some quick searches this seems to be an Intel GPIO controller.
>>   >
>>   > Andy,
>>   >
>>   > Any ideas how to debug next?
>>
>> So what?
>>
>> I will repeat: similar bugs were already fixed here:
>>
>> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/ 
>> commit/?id=782eea0c89f7d071d6b56ecfa1b8b0c81164b9be
>> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/ 
>> commit/?id=a69982c37cd0586e6832268155349301b87f2e35
>> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/ 
>> commit/?id=4cb786180dfb5258ff3111181b5e4ecb1d4a297b
>>
>> So we just need to add similar patch. I. e. something similar to this:
>>
>> +        .matches = {
>> +            DMI_MATCH(DMI_BOARD_NAME, "NL5xRU"),
>> +        },
>> +        .driver_data = &(struct acpi_gpiolib_dmi_quirk) {
>> +            .ignore_wake = "ELAN0415:00@9",
>> +        },
>>
>> Remaining question: how should I get these strings on my system? 
>> "NL5xRU" and "ELAN0415:00@9"?
> 
> DMI_MATCH(DMI_BOARD_NAME, *) will match to /sys/class/dmi/id/board_name 
> (/sys/class/dmi/id/ also contains a lot of other strings to identify 
> your device e.g. /sys/class/dmi/id/sys_vendor which can be matched with 
> DMI_MATCH(DMI_SYS_VENDOR, *))
> 
> How to get "ELAN0415:00@9" is somewhere buried in here https:// 
> gitlab.freedesktop.org/drm/amd/-/issues/1722#note_1725104
> 
> Best regards,
> 
> Werner
> 
>> -- 
>> Askar Safin
>> https://types.pl/@safinaskar
>>

Before jumping into quirk development I think we need to know exactly 
what GPIO is being tripped and that needs to be matched against the Dell 
design (or Intel specs if it's an "internal" GPIO).

That's why I was hoping Andy could help advise.

On the issue that was linked there we knew the GPIO was tied to the 
attention interrupt for touchpad/touchscreen causing the issue so it was 
easy to look at an acpidump and figure out what it's supposed to be.