Laptop wakes up immediately after "suspend" (Acer Nitro ANV14-61 / AMD Ryzen 7 8845HS)
Brief summary of the problem:
The laptop immediately resumes upon entering "suspend" mode with systemctl or by closing the lid.
The problem occurs every time I try to suspend the system on Linux.
All wakeup-related options are disabled in the BIOS. This laptop has a NVIDIA dGPU that is blacklisted in udev (see attached file, but the problems occurs even if not).
According to amd_s2idle, the laptop is awakened by a GPIO event.
Perhaps the only interesting point: this laptop has a "MUX switch" ([Optimus] / dGPU only), and it is the first thing to light up, before the screen. I've attached a video to show this. Maybe this is related to GPIO 11 reported in logs? Unfortunately, at my current level of knowledge, I'm not sure how to investigate this further.
Hardware description:
- amd_s2idle report: s2idle_report-2025-02-09.txt
System information:
- Distro name and Version: Arch Linux x86_64
- Kernel version:
Linux nitro 6.13.2-arch1-1 #1 SMP PREEMPT_DYNAMIC Sat, 08 Feb 2025 18:54:55 +0000 x86_64 GNU/Linux
How to reproduce the issue:
Run sudo systemctl suspend.
Attached files:
Activity
added Phoenix / Hawk Point s0ix labels
I would say GPIO 11 is a pretty good suspicion at the reason. Can you do an experiment to ignore it? There are two tests to do, please do them separately.
- Block the wake source but leave the GPIO enabled. Add this to your kernel command line:
gpiolib_acpi.ignore_wake=AMDI0030:00@11. Run the script again, and see if it still happens. - Block the interrupt for the GPIO entirely. Add this to your kernel command line:
gpiolib_acpi.ignore_interrupt=AMDI0030:00@11. Run the script again, and see if it still happens.
Share the reports from both cases and we can figure out the next steps.
Edited by Mario Limonciello- Block the wake source but leave the GPIO enabled. Add this to your kernel command line:
Thanks a lot for your guidance, @superm1.
Both tests seem to have succeeded this time. The laptop didn't wake up prematurely. Closing the lid also seems to work well now.
Edited by DelganI still don't know what GPIO 11 is connected to, but based on the wakeup I'm suspect it's related to whatever is downstream from this device:
2025-02-11 00:38:14,329 INFO: 🥱 Wakeup triggered from IRQ 38: Advanced Micro Devices, Inc. [AMD] Bridge (0000:00:01.1) (PCIe PME,pciehp,PCIe bwctrl)How exactly are you disabling NVIDIA? Seeing this makes me suspect it should be downstream from
00:01.1:2025-02-11 00:38:14,329 DEBUG: input: HDA NVidia HDMI/DP,pcm=3 as /devices/pci0000:00/0000:00:01.1/0000:01:00.1/sound/card2/input17 2025-02-11 00:38:14,329 DEBUG: input: HDA NVidia HDMI/DP,pcm=7 as /devices/pci0000:00/0000:00:01.1/0000:01:00.1/sound/card2/input18 2025-02-11 00:38:14,329 DEBUG: input: HDA NVidia HDMI/DP,pcm=8 as /devices/pci0000:00/0000:00:01.1/0000:01:00.1/sound/card2/input19 2025-02-11 00:38:14,329 DEBUG: input: HDA NVidia HDMI/DP,pcm=9 as /devices/pci0000:00/0000:00:01.1/0000:01:00.1/sound/card2/input20But I don't see anything connected in the PCI enumeration.
2025-02-11 00:37:56,299 DEBUG: │ 0000:00:01.1 : Advanced Micro Devices, Inc. [AMD] PCI bridge [1022:14ed] : \_SB_.PCI0.GPP0Can you please undo whatever you're doing to disable NVIDIA and let it stay enabled and get the s2idle report with it?
How exactly are you disabling NVIDIA?
I had simply created
/etc/udev/rules.d/00-remove-nvidia.rulesfile with the content attached in the original post of this thread.Reverting this change, NVIDIA components appear in the PCI tree:
2025-02-11 10:54:35,924 DEBUG: │ 0000:00:01.1 : Advanced Micro Devices, Inc. [AMD] PCI bridge [1022:14ed] : \_SB_.PCI0.GPP0 2025-02-11 10:54:35,924 DEBUG: ├─ 0000:01:00.0 : NVIDIA Corporation VGA compatible controller [10de:28a1] : \_SB_.PCI0.GPP0.PEGP 2025-02-11 10:54:35,924 DEBUG: ├─ 0000:01:00.1 : NVIDIA Corporation Audio device [10de:22be] : \_SB_.PCI0.GPP0.NHDAThe test fails if GPIO 11 is not disabled (no change).
s2idle_report-2025-02-11_enable_nvidia_enable_gpio11.txt
s2idle_report-2025-02-11_enable_nvidia_ignore_wake_gpio11.txt
Thanks, with NVIDIA not-disabled the wakeup looks more like what I expect for this issue.
I'm pretty sure this is a BIOS bug that this GPIO is in the ACPI table in the first place. Are you on the latest BIOS? If not; can you please upgrade to the latest one and see if it still happens?
If it still happens:
- Can you please report a bug to Acer.
- Can you please provide your
dmidecodeoutput and I'll come up with a patch that can quirk this by default for your system.
Are you on the latest BIOS?
Yes, this is unfortunately the most recent BIOS version available. I will report this problem to Acer support as you suggested.
Can you please provide your
dmidecodeoutput and I'll come up with a patch that can quirk this by default for your system.Thank you. I attached the
dmidecodeoutput to this message. Out of curiosity, I've also run theamd_s2idle.pyscript on a Ubuntu 24.10 installation, but the behavior is the same (except for the kernel being tainted due to the out-of-tree nvidia drivers). I still attached the reports for completeness.Edited by DelganOK, here you go.
0001-gpiolib-acpi-Add-a-quirk-for-Acer-Nitro-ANV14.patch
If that works, can I please have your name and email address to attribute you? You would be CC'ed on the patch when it's submitted to the mailing list and if it ever needed to be reverted for any reason you would be contacted to test a revert. In this format please.
Reported-and-tested-by: Foo Bar <foo@bar.com>Edited by Mario LimoncielloConfirmed working! Thanks a lot for your exceptional support.
The report and testing can be attributed as follows:
Reported-and-tested-by: Delgan <delgan.py@gmail.com>mentioned in commit FireBurn/linux@8743d669
mentioned in commit nouveau@4e3a6042
mentioned in commit nouveau@9f602c27
mentioned in commit nouveau@bc88b0f9
mentioned in commit nouveau@98fde755
