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

[Andy Shevchenko <andriy.shevchenko@linux.intel.com>]

On Thu, Sep 18, 2025 at 09:33:36PM +0300, Askar Safin wrote:
> Andy Shevchenko:
> > In other words we need to enable debug of the pin control subsystem and see
> > what it will print in dmesg.
> 
> You mean I should enable CONFIG_DEBUG_PINCTRL? Okay, I did this.
> 
> So, today I tested everything on fresh kernel, 6.17.0-rc6, without any
> patches.
> 
> My config is: https://zerobin.net/?ebecc538f6caa22b#88c2k08G8+cZoMjgU9N/WYy28qQjyBW+/H78ygujZxY=
> It was generated from Debian config using localmodconfig.
> I added few tweaks, in particular I enabled CONFIG_DEBUG_PINCTRL.
> 
> # cat /proc/cmdline 
> BOOT_IMAGE=/@rootfs/boot/vmlinuz-6.17.0-rc6 root=UUID=015793d4-ad51-4da7-844b-fcc3bcb13a0b ro rootflags=subvol=@rootfs log_buf_len=4M ignore_loglevel
> 
> I run this script:
> https://zerobin.net/?327f3aa3ef7ce845#Ycu017J9YbRga8uGaCKRzsH7J/lB8D4RudpwTll5lbo=
> 
> This script runs "rtcwake -s 6 -m mem" multiple times. Sometimes my laptop wakes on timer (because of rtcwake),
> and sometimes it wakes up too early on its own (and this is a bug).
> 
> My script did suspend 7 times:
> 
> # dmesg | grep s2idle
> [  117.934504] PM: suspend entry (s2idle)
> [  127.141741] PM: suspend entry (s2idle)
> [  131.299554] PM: suspend entry (s2idle)
> [  140.034802] PM: suspend entry (s2idle)
> [  144.592260] PM: suspend entry (s2idle)
> [  154.038621] PM: suspend entry (s2idle)
> [  163.034299] PM: suspend entry (s2idle)
> 
> Out of them my laptop woke up on timer 4 times and on its own (i. e. due to bug) 3 times:
> 
> # dmesg | grep 'woke up'
> [  126.087936] will-wake: attempt 0: woke up in time
> [  130.248820] will-wake: attempt 1: woke up early
> [  138.988770] will-wake: attempt 2: woke up in time
> [  143.545973] will-wake: attempt 3: woke up early
> [  152.993654] will-wake: attempt 4: woke up in time
> [  161.988956] will-wake: attempt 5: woke up in time
> [  166.329080] will-wake: attempt 6: woke up early
> 
> Here is full output of "dmesg --level=debug+":
> https://zerobin.net/?f704a2d56603f4ec#SRrzc2mt2FNNqcltx/ULmtLZRdRH9frdgoODU03AXwE=
> 
> /proc/interrupts:
> https://zerobin.net/?b7ba5047ca84ab29#TjMUjkAdhpIuKbnvPpuYyNWa/ilA/ciGKwwSbx6KRFc=


  14:          0          0          0          0          0          0          0          0          0          0          0          0          0          0          0          0          0          0          0          0       4202          0          0          0          0          0          0          0          0          0          0          0 IR-IO-APIC   14-fasteoi   INTC1085:00


 204:          0          0          0          0          0          0          0          0          0          0          0          0          0          0          0          0          0          0          0          0       4199          0          0          0          0          0          0          0          0          0          0          0 intel-gpio  355  VEN_0488:00


Sounds like it comes via GPIO, but it's not handled as touchpad IRQ. You may
try to add a quirk to prevent touchpad IRQ from waking the system. That should
help I believe.
Something like "ignore_wake=INTC1085:00@355" in the kernel command line.
If it helps, update drivers/gpio/gpiolib-acpi-quirks.c accordingly.

-- 
With Best Regards,
Andy Shevchenko