Re: [BUG]i2c_hid_acpi broken with 4.17.2 on Framework Laptop 13 AMD

[Sébastien Chaumat <euidzero@xxxxxxxxx>]

Any direction on how I can enhance the debugging at the kernel level ?

There was an old issue with amd_gpio there : https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1971597

Coud the kernel be confused by IRQ/GSI mapping ? Any way to test this hypothesis?

Thanks

Le mar. 5 déc. 2023 à 09:17, Jan Beulich <jbeulich@xxxxxxxx> a écrit :

On 04.12.2023 20:17, Sébastien Chaumat wrote:
> Le lun. 4 déc. 2023 à 10:06, Jan Beulich <jbeulich@xxxxxxxx> a écrit :
>
>> On 03.12.2023 10:56, Sébastien Chaumat wrote:
>>> Hello,
>>>
>>>  Trying to get the Framework Laptop 13 AMD to work with QubesOS I hit the
>>> following Xen issue :
>>>
>>> Xen version : 4.17.2
>>
> + tested with 4.18.0
>
>
>>> Kernel : 6.5.12-300.fc39.x86_64
>>> CPU  model name : AMD Ryzen 7 7840U w/ Radeon  780M Graphics
>>
>
>
>>> [    2.464598] amd_gpio AMDI0030:00: failed to enable wake-up interrupt
>>
>> Possibly releated to this. You'll want to obtain a full-verbosity
>> hypervisor
>> log with a debug hypervisor, as there may be hypervisor debug messages
>> telling us what Xen may not like.
>>
>
> xl dmesg with some traces attached.

Nothing that looks relevant here. The anomalies are related to you also
having enabled lock debugging and ubsan.

I'm afraid this needs looking into from the kernel side first, to understand
what's going wrong there. Once that's known, it'll hopefully be more clear
whether this is a Xen or a kernel issue.

Jan