Re: [PATCH 2/4] ACPI: button: Cancel hibernation if button is pressed during hibernation

[Muhammad Usama Anjum <usama.anjum@collabora.com>]

On 11/26/25 5:55 PM, Mario Limonciello wrote:
>>> First I need a board/machine with serial console access to view all logs in real
>>> time. :)
>>
>> usb debug cables might be your solution.
>>
> Just a word of warning before you go too far down this path to get a console working from XHCI debug.
> 
> This is probably a Hal changing a light bulb problem [1]. Last time I tried XHCI debug cables on some modern AMD systems I ran into a problem that the BAR is too big for early_ioremap().  Looking through LKML it's come up a few times in the past too [2] [3].
> 
> Link: https://youtu.be/5W4NFcamRhM?si=qOFrCTzvK6-H-4AX [1]
> Link: https://lore.kernel.org/linux-usb/ZCOq3PUBCtHkwdnw@mail-itl/ [2]
> Link: https://lore.kernel.org/linux-usb/CAAcb1K_bezseTM8DrOrzVUi_W+nZoE2N0CO4k3AQWPw7=7pyjw@mail.gmail.com/ [3]
> 
> The other obvious idea is to use netconsole, but you need a PCIe Ethernet controller, but I think you'll have more success in development using KVM as these are generic architectural problems.
> 
> To help you get started with this I may point out something that was shared to me for another hibernate bug [4].
> 
> Link: https://lore.kernel.org/linux-pm/20251105180506.137448-1-safinaskar@gmail.com/ [4]
> 
> Askar Safin (CC'ed) produced a script that does a very minimal kernel build, sets up a VM with the right sizes of disks/swap/etc.  It's trivial to make kernel changes and re-run the script, and you can  also attach a debugger to the KVM instance.  Maybe you can adapt something like that.  You can wrap it with 'time' calls to actually measure performance for any ideas and prove them out too.
I was just going to try it. Thank you so much for double confirming. I'll
test and see.

-- 
---
Thanks,
Usama