[SOLVED] Can't resume from suspend ....

saronno · 2023-08-23 14:13:29

I just move my installation on a new pc ... everything is fine but resume-suspend. Everytime I try to suspend the system I just can't resume it.

How can I debug this issue?

It's probably an issue with the new hardware, maybe the video card (an RTX 3050). The mainboard ACPI string is "windows 2013" .

I try to use systemd nvidia interface (the scripts they release ... I am using the proprietary drivers), same thing.

Any idea?

TIA

Last edited by saronno (2023-09-27 12:42:56)

saronno · 2023-08-24 07:45:55

logs

http://ix.io/4Env

seth · 2023-08-24 08:00:08

You've NM and systemd-networkd enabled in parallel, pick one, disable the other.

The mainboard ACPI string is "windows 2013" .

There's no parallel windows installation, is there?

It's probably an issue with the new hardware, maybe the video card (an RTX 3050).

https://download.nvidia.com/XFree86/Lin … ement.html

Aug 23 15:40:50 kanon kernel: [drm] Initialized simpledrm 1.0.0 20200625 for simple-framebuffer.0 on minor 0

https://wiki.archlinux.org/title/NVIDIA … de_setting

Everytime I try to suspend the system I just can't resume it.

What exactly happens instead?

saronno · 2023-08-24 09:05:09

seth wrote:

You've NM and systemd-networkd enabled in parallel, pick one, disable the other.

done, ty for the advise.

There's no parallel windows installation, is there?

No, only arch. The only thing I add is that I tested the system with a disk with windows 10 installed before moving my arch linux in with the new nvme ssd.
Could this interfere with the BIOS ACPI setup?

https://download.nvidia.com/XFree86/Lin … ement.html

I will try some hints I found there and report back ...

Aug 23 15:40:50 kanon kernel: [drm] Initialized simpledrm 1.0.0 20200625 for simple-framebuffer.0 on minor 0

https://wiki.archlinux.org/title/NVIDIA … de_setting

Done

Aug 24 07:31:44 kanon kernel: [drm] [nvidia-drm] [GPU ID 0x00000300] Loading driver
Aug 24 07:31:45 kanon kernel: [drm] Initialized nvidia-drm 0.0.0 20160202 for 0000:03:00.0 on minor 0

What exactly happens instead?

System power up, no signal whatsover on the screens (they remain inactive), there is some disks activity; after a while system seem to start the resume process all over again(for a second it acts as it power down and then it power up again with all the fans and led on .. and a bit of disk activity). In other words, no signal on screen and system doesn't stay on but try to repeat the cycle every 20-25 seconds.

Should I keep this configuration?

[saronno@kanon ~]$ cat /sys/power/mem_sleep
s2idle [deep]

Last edited by saronno (2023-08-24 09:12:29)

saronno · 2023-08-24 12:12:41

I add also xorg log because it's very messy.

http://ix.io/4Eos

It gives back error loading module and also something about acpi who could be interesting for the sleep-resume problem.

[saronno@kanon ~]$ pacman -Qs nouveau
[saronno@kanon ~]$ pacman -Qs xf86-video
local/xf86-video-amdgpu 23.0.0-1 (xorg-drivers)
    X.org amdgpu video driver
local/xf86-video-ati 1:22.0.0-1 (xorg-drivers)
    X.org ati video driver
local/xf86-video-fbdev 0.5.0-3 (xorg-drivers)
    X.org framebuffer video driver
local/xf86-video-vesa 2.6.0-1 (xorg-drivers xorg)
    X.org vesa video driver
[saronno@kanon ~]$ pacman -Qs nvidia
local/egl-wayland 2:1.1.12-1
    EGLStream-based Wayland external platform
local/lib32-nvidia-utils 535.98-1
    NVIDIA drivers utilities (32-bit)
local/libvdpau 1.5-1
    Nvidia VDPAU library
local/libxnvctrl 535.98-1
    NVIDIA NV-CONTROL X extension
local/nvidia 535.98-4
    NVIDIA drivers for linux
local/nvidia-settings 535.98-1
    Tool for configuring the NVIDIA graphics driver
local/nvidia-utils 535.98-1
    NVIDIA drivers utilities

seth · 2023-08-24 13:09:17

[    34.484] (EE) Failed to load module "nouveau" (module does not exist, 0)
[    34.484] (EE) Failed to load module "nv" (module does not exist, 0)

is perfectly fine, you don't want to use those drivers; the xorg log doesn't suggest any problems.

s2idle will probably not cause the same problems but also limit the power saving during the "sleep"
b/c of "mem_sleep_default=deep ": does the system default to s2idle?

System power up, no signal whatsover on the screens (they remain inactive), there is some disks activity; after a while system seem to start the resume process all over again(for a second it acts as it power down and then it power up again with all the fans and led on .. and a bit of disk activity). In other words, no signal on screen and system doesn't stay on but try to repeat the cycle every 20-25 seconds.

This here sounds like sounds a bit like the system is underpowered, can you https://wiki.archlinux.org/title/Keyboa … el_(SysRq) during the wakeup?
Can you reliably sleep and wake from the multi-user.target (no graphical target running at all)?
Does the problem depend on the sleep duration? (Ie. can you wake the system after 5s or so?)

pacman -Qs xf86-video

You don't need nor want any of those, but they're not the problem.

PSA:

[    34.983] (**) Microsoft Comfort Curve Keyboard 2000: Applying InputClass "evdev keyboard catchall"

If it ever acts up and some keys don't work or work infrequently, particularily w/ humidity changes: you've tho whack it (seriously, grab it by the top near the cable an give it a firm spanking ;-)

saronno · 2023-08-24 14:59:42

This here sounds like sounds a bit like the system is underpowered, can you https://wiki.archlinux.org/title/Keyboa … el_(SysRq) during the wakeup?
Can you reliably sleep and wake from the multi-user.target (no graphical target running at all)?
Does the problem depend on the sleep duration? (Ie. can you wake the system after 5s or so?)

Problem doesn't depend on duration.
If I boot to terminal only the problem is still there, same behaviour trying to resume after suspend.
I tried to use sysreq shortcuts after enabling them: no way. Completely ignored them. Of course I tried reboot ... but it keeps looping as describe above.
Sysreq shortcuts works fine if I use them when the machine is on.

PSA:
[    34.983] (**) Microsoft Comfort Curve Keyboard 2000: Applying InputClass "evdev keyboard catchall"
If it ever acts up and some keys don't work or work infrequently, particularily w/ humidity changes: you've tho whack it (seriously, grab it by the top near the cable an give it a firm spanking ;-)

I have no problem with it ... anyway I spank it regularly after put it upside down for cleaning :-)

Last edited by saronno (2023-08-24 15:40:04)

seth · 2023-08-24 18:38:09

https://wiki.archlinux.org/title/Power_ … nsistently
https://01.org/blogs/rzhang/2015/best-p … ate-issues

seth wrote:

s2idle will probably not cause the same problems but also limit the power saving during the "sleep"
b/c of "mem_sleep_default=deep ": does the system default to s2idle?

saronno · 2023-08-24 21:53:40

I added mem_sleep_default=deep to the boot line ...

[saronno@kanon ~]$ cat /sys/power/mem_sleep
s2idle [deep]

Nothing change.

seth · 2023-08-24 21:55:29

The parameter was already there what made me wonder whether you added it there because it would default to s2idle
You can also just "echo s2idle | sudo tee /sys/power/mem_sleep" to transiently chage the mode.

saronno · 2023-08-25 06:23:58

What about all that [Firmware Bug] I get from the kernel .... is safe to ignore them?

seth · 2023-08-25 06:51:07

The only interesting one is

Aug 23 15:40:50 kanon kernel: ACPI: [Firmware Bug]: BIOS _OSI(Linux) query ignored

I might have mistinterpreted

The mainboard ACPI string is "windows 2013" .

Did you actually try

acpi_osi=! acpi_osi="Windows 2013"

https://learn.microsoft.com/en-us/windo … inacpi-osi

saronno · 2023-08-25 07:50:07

Yes, I did, but nothing changed in the kernel log.

saronno · 2023-08-25 08:06:31

seth wrote:

The parameter was already there what made me wonder whether you added it there because it would default to s2idle
You can also just "echo s2idle | sudo tee /sys/power/mem_sleep" to transiently chage the mode.

Anyway you were right, s2idle works fine.

This, maybe, should help to pinpoint where the problem is .....

saronno · 2023-08-25 20:48:21

Some interesting news ....

I try to downgrade the BIOS (theprevious one was the latest, but it was also in beta) .... same story.

I boot from the windows installation I use to test teh system after the assembling; I use a usb 3 external case ... and I try to suspend the system and resume: same behaviour.

So, it seem there is something wrong with BIOS, hardware or one of the firmware of the devices installed.

seth · 2023-08-26 13:44:35

You could try https://wiki.archlinux.org/title/Solid_ … nd_support

saronno · 2023-08-27 09:33:25

seth wrote:

You could try https://wiki.archlinux.org/title/Solid_ … nd_support

I tried it, same story.

So, I tried to install another video card .... no way. So I tried to remove the nvme drive .... still same story, even under windows.

To me it sound like a bios bug. And since I tried the last two BIOS release, I can't do anything more.

Finally, yesterday I tried also to update microcode .... this result in my system completely mess up: I got a kernel panic .... then I remove the string from kernel parameters, kernel panic again ... I try with a live cd ..kernel panic.
Then I reset the cmos ... kernel panic. Finally I tried to boot from windows .... it works. The I tried again linux ...and It worked.

My system is based on a x99s sli plus board (from MSI) and an i7-5820k ....
anyway, a part from this things, the system is stable and can stay up without any issue for days.

PS: the microcode update mess was my fault, I give it the wrong initrd path because I forgot I have a separate boot partition.

Last edited by saronno (2023-08-27 19:34:14)

saronno · 2023-08-30 13:35:36

I replaced the PSU with a known good (it was a known good also the previous one) ..... same story.

At this point, I have no idea how to solve this.

Last edited by saronno (2023-08-30 17:19:24)

seth · 2023-09-12 13:03:01

https://web.archive.org/web/20230204052 … ate-issues
But iff the system defaults to s2idle that might be the only reliable sleep variant.

saronno · 2023-09-12 14:19:06

As you can see in msg #4, the default was deep, not s2idle. I added deep to the kernel line just to be sure at the beginning. Now I added s2idle to the kernel line to make it default since it's the only one working.
Anyway, since I can exclude the nvme disk, the mouse and the graphics card, I think it's a motherboard/firmware fault. The problem is not present only in linux, but also in windows (it's not installed in the system, but I use a disk with windows just to try ...) and other live cd.
It goes asleep, then try to wake up, but if fails resulting in a costant loop of 10-15 second where it try again to wake everything up from start.

I have not already check, but I will be not surprise if memory lacks voltage during sleep phase. This will explain the behaviour, since it will not find anything in ram when you try to wake it up.

Magnesium · 2023-09-19 12:29:17

Use downgrade package to go back to kernel 6.4.11.arch2-1 OR maybe 6.4.12 seems ok

Here the resuming from suspension is working again.

best regards.

seth · 2023-09-19 20:55:04

(it's not installed in the system, but I use a disk with windows just to try ...)

For clarification: if you're juggling linux and windows disks, that amounts to dual-booting and the 3rd link below becomes relevant ("very" in this specific context)

Problem doesn't depend on duration.
If I boot to terminal only the problem is still there

makes RAM issues less likely, you'd effectively have to lose all current immediately.
Does your BIOS allow altering the RAM config (voltage, clock, timings)?

saronno · 2023-09-20 22:30:17

seth wrote:

(it's not installed in the system, but I use a disk with windows just to try ...)
For clarification: if you're juggling linux and windows disks, that amounts to dual-booting and the 3rd link below becomes relevant ("very" in this specific context)
Problem doesn't depend on duration.
If I boot to terminal only the problem is still there
makes RAM issues less likely, you'd effectively have to lose all current immediately.
Does your BIOS allow altering the RAM config (voltage, clock, timings)?

Yes, the bios allow to setup voltage, clock and timings for rams. (actually it has xpm options).
Anyway, everything is set in auto mode so, cpu and ram parameters are the default ones.

saronno · 2023-09-20 22:33:04

Magnesium wrote:

Use downgrade package to go back to kernel 6.4.11.arch2-1 OR maybe 6.4.12 seems ok
Here the resuming from suspension is working again.
best regards.

So, you had problems to wake up the pc from suspension with the latest kernels?

I don't know, because I have the same problem even with a live cd from ubuntu or manjaro.

seth · 2023-09-21 07:01:57

Anyway, everything is set in auto mode so, cpu and ram parameters are the default ones.

The idea would be to turn the RAM more conservative than the default (in case it truely is unreliable)

What board exactly is it?
Also

dmidecode

Arch Linux

#1 2023-08-23 14:13:29

[SOLVED] Can't resume from suspend ....

#2 2023-08-24 07:45:55

Re: [SOLVED] Can't resume from suspend ....

#3 2023-08-24 08:00:08

Re: [SOLVED] Can't resume from suspend ....

#4 2023-08-24 09:05:09

Re: [SOLVED] Can't resume from suspend ....

#5 2023-08-24 12:12:41

Re: [SOLVED] Can't resume from suspend ....

#6 2023-08-24 13:09:17

Re: [SOLVED] Can't resume from suspend ....

#7 2023-08-24 14:59:42

Re: [SOLVED] Can't resume from suspend ....

#8 2023-08-24 18:38:09

Re: [SOLVED] Can't resume from suspend ....

#9 2023-08-24 21:53:40

Re: [SOLVED] Can't resume from suspend ....

#10 2023-08-24 21:55:29

Re: [SOLVED] Can't resume from suspend ....

#11 2023-08-25 06:23:58

Re: [SOLVED] Can't resume from suspend ....

#12 2023-08-25 06:51:07

Re: [SOLVED] Can't resume from suspend ....

#13 2023-08-25 07:50:07

Re: [SOLVED] Can't resume from suspend ....

#14 2023-08-25 08:06:31

Re: [SOLVED] Can't resume from suspend ....

#15 2023-08-25 20:48:21

Re: [SOLVED] Can't resume from suspend ....

#16 2023-08-26 13:44:35

Re: [SOLVED] Can't resume from suspend ....

#17 2023-08-27 09:33:25

Re: [SOLVED] Can't resume from suspend ....

#18 2023-08-30 13:35:36

Re: [SOLVED] Can't resume from suspend ....

#19 2023-09-12 13:03:01

Re: [SOLVED] Can't resume from suspend ....

#20 2023-09-12 14:19:06

Re: [SOLVED] Can't resume from suspend ....

#21 2023-09-19 12:29:17

Re: [SOLVED] Can't resume from suspend ....

#22 2023-09-19 20:55:04

Re: [SOLVED] Can't resume from suspend ....

#23 2023-09-20 22:30:17

Re: [SOLVED] Can't resume from suspend ....

#24 2023-09-20 22:33:04

Re: [SOLVED] Can't resume from suspend ....

#25 2023-09-21 07:01:57

Re: [SOLVED] Can't resume from suspend ....

Board footer