PC immediately wakes up after suspend to RAM

magicgoose · 2022-06-12 12:52:55

Hi. Suspend to RAM used to work fine for a long time. Then, after the PC sustained some extra physical wear and tear, and after I had to replace both keyboard and mouse, it fails to correctly suspend to RAM (it goes into sleep fine, but almost immediately wakes up). Still using the same Arch Linux setup as before, didn't reinstall.
Maybe this could be related to mouse and/or keyboard. Sometimes if I disconnect them both before suspending, it might stay asleep, but this is not guaranteed. Maybe it was just a coincidence.
I have Windows 10 on another disk (just to play a game occasionally), and with Windows this problem doesn't happen, and I don't have to disconnect any devices.
What can I do to make the sleep work correctly on Arch Linux as well?

More details:
Operating System: Arch Linux
KDE Plasma Version: 5.24.5
KDE Frameworks Version: 5.94.0
Qt Version: 5.15.4
Kernel Version: 5.18.3-zen1-1-zen (64-bit)
Graphics Platform: X11
Processors: 16 × AMD Ryzen 7 3800X 8-Core Processor
Memory: 62.7 GiB of RAM
Graphics Processor: AMD Radeon RX 590 Series

Both keyboard and mouse are wireless, with USB dongles, made by Logitech. Previously (when everything worked fine) I had a different wired keyboard and wireless mouse, all also by Logitech.

Last edited by magicgoose (2022-06-12 13:18:05)

Maniaxx · 2022-06-12 13:25:39

Disable them one by one to get the bad device.
https://wiki.archlinux.org/title/Wakeup … om_suspend

If both are faulty disable them in BIOS and use the power button.

magicgoose · 2022-06-12 14:18:04

Thanks for pointer. I looked at /proc/acpi/wakeup, there were 6 enabled entries. I disabled them all:

After this, all entries in /proc/acpi/wakeup are disabled. Then I tried to suspend and the same problem still happens.
I'll try to see what I can do in the PC motherboard's firmware menu (BIOS) in a few minutes...

magicgoose · 2022-06-12 14:28:58

Unfortunately looks like everything that's relevant is already disabled in BIOS. I can upload a photo if anyone's interested how it looks, but there are settings named:

"Power On By Keyboard"
"Power On By Mouse"
"Resume by Alarm"
"Wake on LAN"
and they are all disabled. There are also a couple other settings on that screen which I don't recognize, some cryptic words.

seth · 2022-06-12 14:31:32

I have Windows 10 on another disk

3rd link below - not a guarantee, but the ACPI in most affected by this. Reboot windows and linux twice afterwards (for voodoo reasons)

magicgoose · 2022-06-12 15:19:15

Looks like hibernation was already disabled in this Windows installation. Just in case, I did it anyway -- ran "powercfg -h off" from admin powershell window. Then shut down. After that, the problem still happens. I saw it needs to be done twice, I'll check later again.

Last edited by magicgoose (2022-06-12 15:20:01)

Maniaxx · 2022-06-12 15:29:44

Try this after it woke up by itself (to show wakeup reason). Maybe it delivers a hint or something.

$ sudo dmidecode | grep Wake-up

magicgoose · 2022-06-12 20:35:39

After turning on, no sleep just yet:
~> sudo dmidecode | grep Wake-up
Wake-up Type: Power Switch

Then, after sleep (and unwanted immediate wake-up):
~> sudo dmidecode | grep Wake-up
Wake-up Type: PCI PME#

Searched this a bit, looks like this is used for the keyboard, but I guess that's just one of the many possibilities.

seth · 2022-06-12 21:03:02

You could try passing

pcie_port_pm=off pcie_pme=nomsi

to the kernel.
Do you still have a wired keyboard?

Maniaxx · 2022-06-13 13:48:53

Maybe raising the ACPI debug verbosity leads to something.

Enable all ACPI hardware-related messages:
$ sudo sh -c "echo 0x2 > /sys/module/acpi/parameters/debug_layer"
$ sudo sh -c "echo 0xffffffff > /sys/module/acpi/parameters/debug_level"

https://www.kernel.org/doc/Documentatio … /debug.rst

magicgoose · 2022-08-21 09:25:27

Thanks.
I don't have my previous wired keyboard, and looks like my current wireless one doesn't have a wired connection option - the USB port is only used for charging and maybe for firmware updates, but it doesn't work as a keyboard when connected by wire.

In the 1st ACPI command above, I tried to understand mask by reading your link, and it means only ACPI_HARDWARE and nothing else from "supported mask values"?
Where should I look for the written log, journalctl?

Maniaxx · 2022-08-21 12:05:20

magicgoose wrote:

In the 1st ACPI command above, I tried to understand mask by reading your link, and it means only ACPI_HARDWARE and nothing else from "supported mask values"?
Where should I look for the written log, journalctl?

Yes and yes.

magicgoose · 2022-08-21 12:38:43

Ran these 2 commands, then put it to sleep, and then looked at the output of
journalctl -r

There is a lot of text related to ACPI, unsure what is the important part. It's also all posted with "kernel" tag.
Is there a simple way to filter out unrelated messages which are also under "kernel"?

seth · 2022-08-21 13:04:06

You can esp. filter by the boot "sudo journalctl -b"

Maniaxx · 2022-08-21 15:28:41

Kernel buffer only:

$ sudo journalctl -b --dmesg

$ sudo dmesg --color=always |less

You can filter with grep, awk, etc. but i don't think its necessary/useful. The other kernel messages might help for orientation (system state, etc).
Maybe clear the dmesg buffer first.

$ sudo dmesg -C

Its possible that 'dmesg' hits the kernel ring buffer limit so it might be better to use 'journalctl' instead (or raise the ring buffer).

You can also limit the ACPI debug verbosity (if you know what you're doing).

$ cat /sys/module/acpi/parameters/debug_level
Description              	Hex        SET
ACPI_LV_INIT             	0x00000001 [ ]
ACPI_LV_DEBUG_OBJECT     	0x00000002 [ ]
ACPI_LV_INFO             	0x00000004 [*]
ACPI_LV_REPAIR           	0x00000008 [*]
ACPI_LV_TRACE_POINT      	0x00000010 [ ]
ACPI_LV_INIT_NAMES       	0x00000020 [ ]
ACPI_LV_PARSE            	0x00000040 [ ]
ACPI_LV_LOAD             	0x00000080 [ ]
ACPI_LV_DISPATCH         	0x00000100 [ ]
ACPI_LV_EXEC             	0x00000200 [ ]
ACPI_LV_NAMES            	0x00000400 [ ]
ACPI_LV_OPREGION         	0x00000800 [ ]
ACPI_LV_BFIELD           	0x00001000 [ ]
ACPI_LV_TABLES           	0x00002000 [ ]
ACPI_LV_VALUES           	0x00004000 [ ]
ACPI_LV_OBJECTS          	0x00008000 [ ]
ACPI_LV_RESOURCES        	0x00010000 [ ]
ACPI_LV_USER_REQUESTS    	0x00020000 [ ]
ACPI_LV_PACKAGE          	0x00040000 [ ]
ACPI_LV_ALLOCATIONS      	0x00100000 [ ]
ACPI_LV_FUNCTIONS        	0x00200000 [ ]
ACPI_LV_OPTIMIZATIONS    	0x00400000 [ ]
ACPI_LV_MUTEX            	0x01000000 [ ]
ACPI_LV_THREADS          	0x02000000 [ ]
ACPI_LV_IO               	0x04000000 [ ]
ACPI_LV_INTERRUPTS       	0x08000000 [ ]
ACPI_LV_AML_DISASSEMBLE  	0x10000000 [ ]
ACPI_LV_VERBOSE_INFO     	0x20000000 [ ]
ACPI_LV_FULL_TABLES      	0x40000000 [ ]
ACPI_LV_EVENTS           	0x80000000 [ ]
--
debug_level = 0x0000000C (* = enabled)

Last edited by Maniaxx (2022-08-21 15:51:26)

Arch Linux

#1 2022-06-12 12:52:55

PC immediately wakes up after suspend to RAM

#2 2022-06-12 13:25:39

Re: PC immediately wakes up after suspend to RAM

#3 2022-06-12 14:18:04

Re: PC immediately wakes up after suspend to RAM

#4 2022-06-12 14:28:58

Re: PC immediately wakes up after suspend to RAM

#5 2022-06-12 14:31:32

Re: PC immediately wakes up after suspend to RAM

#6 2022-06-12 15:19:15

Re: PC immediately wakes up after suspend to RAM

#7 2022-06-12 15:29:44

Re: PC immediately wakes up after suspend to RAM

#8 2022-06-12 20:35:39

Re: PC immediately wakes up after suspend to RAM

#9 2022-06-12 21:03:02

Re: PC immediately wakes up after suspend to RAM

#10 2022-06-13 13:48:53

Re: PC immediately wakes up after suspend to RAM

#11 2022-08-21 09:25:27

Re: PC immediately wakes up after suspend to RAM

#12 2022-08-21 12:05:20

Re: PC immediately wakes up after suspend to RAM

#13 2022-08-21 12:38:43

Re: PC immediately wakes up after suspend to RAM

#14 2022-08-21 13:04:06

Re: PC immediately wakes up after suspend to RAM

#15 2022-08-21 15:28:41

Re: PC immediately wakes up after suspend to RAM

Board footer