Suspend and resume - black screen

[sebbouckaert]

Hi, since recent updates from the last week or so my computer is giving me a black screen every time I try to resume. This regression occurred since the latest couple of updates including a new kernel version.

System details:

Linux Mint 20.2, Cinnamon 5.05, Kernel version 5.4.0-84-generic, GPU Nvidia Geforce GTX 750 ti.

GPU used to run on driver v. 360. But then I saw v.370 was available (recommended), so I upgraded that first. No improvement.

Then I tried this fix mentioned here for use with v370:

https://forums.developer.nvidia.com/t/f ... 0/187150/3

So I removed the nVidia systemd services mentioned. No improvement, on the contrary: with the 360 driver the funny thing was that when I went into suspend mode manually and pressed the power to resume, everything worked well resulting in login screen. Only when leaving the PC idle for some time (power settings are 1 hour idle before suspend) and then try to resume again, the black screen issue occurred. Since this 'fix' I always get the black screen (monitor HDMI doesn't receive any video signal).

Needless to say how annoying this is. I work with quite a few opened tabs in browser, some of them logged in to various sites, everything needs reopening and logging in after each break. So, any suggestions? Thanks!

[Hoser Rob]

I'm no nvidia expert but I can say that that fix you linked was for the 460/470 nvidia drivers, anf those are not the ones for your card. Be careful with readnjg dev forums. They're going to assume a high skill level on the part of the readers.

And you need to post more info, copy/paste the complete text output of inxi -Fxxxrz.

[sebbouckaert]

Are you sure about your remark concerning these drivers not right for my card? Nvidia-settings clearly states that driver 470.63.01 is actually running whereas 460 did the job the previous months. Mint's Driver Manager only offers the following options: 340, 390, 460, 470 and of course the nouveau driver. That last one is not an option for me, resulting in serious tearing issues all over.

Anyway, here's my output from inxi -Fxxxrz.

Code: Select all

 System:
  Kernel: 5.4.0-84-generic x86_64 bits: 64 compiler: gcc v: 9.3.0 
  Desktop: Cinnamon 5.0.5 wm: muffin 5.0.1 dm: LightDM 1.30.0 
  Distro: Linux Mint 20.2 Uma base: Ubuntu 20.04 focal 
Machine:
  Type: Desktop Mobo: MSI model: A68HM-P33 (MS-7721) v: 8.0 serial: <filter> 
  UEFI [Legacy]: American Megatrends v: 34.2 date: 09/26/2014 
CPU:
  Topology: Quad Core model: AMD Athlon X4 860K bits: 64 type: MCP 
  arch: Steamroller rev: 1 L2 cache: 2048 KiB 
  flags: avx lm nx pae sse sse2 sse3 sse4_1 sse4_2 sse4a ssse3 svm 
  bogomips: 29546 
  Speed: 1696 MHz min/max: 1700/3700 MHz boost: enabled Core speeds (MHz): 
  1: 1927 2: 1846 3: 1694 4: 1696 
Graphics:
  Device-1: NVIDIA GM107 [GeForce GTX 750 Ti] vendor: Micro-Star MSI 
  driver: nvidia v: 470.63.01 bus ID: 01:00.0 chip ID: 10de:1380 
  Display: x11 server: X.Org 1.20.11 driver: nvidia tty: N/A 
  OpenGL: renderer: NVIDIA GeForce GTX 750 Ti/PCIe/SSE2 
  v: 4.6.0 NVIDIA 470.63.01 direct render: Yes 
Audio:
  Device-1: AMD FCH Azalia vendor: Micro-Star MSI driver: snd_hda_intel 
  v: kernel bus ID: 00:14.2 chip ID: 1022:780d 
  Device-2: NVIDIA GM107 High Definition Audio [GeForce 940MX] 
  vendor: Micro-Star MSI driver: snd_hda_intel v: kernel bus ID: 01:00.1 
  chip ID: 10de:0fbc 
  Device-3: Focusrite-Novation Focusrite Scarlett 2i2 type: USB 
  driver: snd-usb-audio bus ID: 2-1:2 chip ID: 1235:8016 
  Sound Server: ALSA v: k5.4.0-84-generic 
Network:
  Device-1: Realtek RTL8111/8168/8411 PCI Express Gigabit Ethernet 
  vendor: Micro-Star MSI driver: r8169 v: kernel port: d000 bus ID: 02:00.0 
  chip ID: 10ec:8168 
  IF: enp2s0 state: down mac: <filter> 
  Device-2: Ralink RT2870/RT3070 Wireless Adapter type: USB 
  driver: rt2800usb bus ID: 2-2:3 chip ID: 148f:3070 serial: <filter> 
  IF: wlx001f1f59e818 state: up mac: <filter> 
Drives:
  Local Storage: total: 2.73 TiB used: 770.38 GiB (27.6%) 
  ID-1: /dev/sda vendor: Samsung model: SSD 860 EVO 500GB size: 465.76 GiB 
  speed: 6.0 Gb/s serial: <filter> rev: 4B6Q scheme: MBR 
  ID-2: /dev/sdb vendor: Western Digital model: WD5000AZRX-00A8LB0 
  size: 465.76 GiB speed: 6.0 Gb/s serial: <filter> rev: 1A01 scheme: MBR 
  ID-3: /dev/sdc type: USB vendor: Western Digital model: WD Elements 2621 
  size: 1.82 TiB serial: <filter> rev: 1026 scheme: GPT 
Partition:
  ID-1: / size: 322.03 GiB used: 131.48 GiB (40.8%) fs: ext4 dev: /dev/sda2 
  ID-2: swap-1 size: 20.21 GiB used: 0 KiB (0.0%) fs: swap dev: /dev/sda3 
Sensors:
  System Temperatures: cpu: 15.5 C mobo: N/A gpu: nvidia temp: 40 C 
  Fan Speeds (RPM): N/A gpu: nvidia fan: 30% 
Repos:
  No active apt repos in: /etc/apt/sources.list 
  Active apt repos in: /etc/apt/sources.list.d/eid.list 
  1: deb http://files.eid.belgium.be/debian ulyana main
  2: deb http://files2.eid.belgium.be/debian ulyana main
  Active apt repos in: /etc/apt/sources.list.d/flacon-ppa-focal.list 
  1: deb http://ppa.launchpad.net/flacon/ppa/ubuntu focal main
  Active apt repos in: /etc/apt/sources.list.d/google-chrome.list 
  1: deb [arch=amd64] http://dl.google.com/linux/chrome/deb/ stable main
  Active apt repos in: /etc/apt/sources.list.d/official-package-repositories.list 
  1: deb http://packages.linuxmint.com uma main upstream import backport #id:linuxmint_main
  2: deb http://archive.ubuntu.com/ubuntu focal main restricted universe multiverse
  3: deb http://archive.ubuntu.com/ubuntu focal-updates main restricted universe multiverse
  4: deb http://archive.ubuntu.com/ubuntu focal-backports main restricted universe multiverse
  5: deb http://security.ubuntu.com/ubuntu/ focal-security main restricted universe multiverse
  6: deb http://archive.canonical.com/ubuntu/ focal partner
  Active apt repos in: /etc/apt/sources.list.d/spotify.list 
  1: deb http://repository.spotify.com stable non-free
Info:
  Processes: 261 Uptime: 6h 46m Memory: 15.65 GiB used: 5.64 GiB (36.1%) 
  Init: systemd v: 245 runlevel: 5 Compilers: gcc: 9.3.0 alt: 9 Shell: bash 
  v: 5.0.17 running in: gnome-terminal inxi: 3.0.38 

[SMG]

Hi, since recent updates from the last week or so my computer is giving me a black screen every time I try to resume. This regression occurred since the latest couple of updates including a new kernel version.

Do you remember if those recent updates included an Nvidia driver update?

If so, do you have any Timeshift snapshots taken while using the older Nvidia driver? Quite a few people have been running into issues related to suspension/power states with recent Nvidia drivers. When you run into problems, it is usually best to roll back to an older snapshot and then apply updates one at a time to see which one is the problem.

Are you sure about your remark concerning these drivers not right for my card? Nvidia-settings clearly states that driver 470.63.01 is actually running whereas 460 did the job the previous months.

Your initial post talked about using the 360 and 370 drivers and that might have caused confusion. The Nvidia-470 is the latest driver available for your card.

I'm not clear on where you stand right now. You were using the Nvidia-460 and it was working before you did the updates, but now it is not? Or was it not working correctly even before you updated to the Nvidia-470?

I understand you say you are having problems with the Nvidia-470, but you do not have to update to it and can stay with the Nvidia-460 if that works.

[sebbouckaert]

1) The update mentioned was only for the Mint system through the usual Mint Update, so no proprietary nvidia drivers since those are offered through the Device driver manager (in Cinnamon). There was a kernel update, however.
2) No recent Timeshift snapshots I'm afraid.
3) Summary of the problem: I was running v460 of the GPU driver (I guess for months without any problem). Then, after the (kernel?) update the issue started. So I updated to 470 hoping to resolve the issue (new kernel, new GPU driver update, sounds logical, and my system mentioned "recommended" with v470. Problem still persisted. Only then, I deleted the nvidia systemd services as mentioned earlier.
Strange thing is: sometimes the computer resumes normally. This mostly when I manually go into suspend mode (menu > shutdown > suspend) and resume by pressing power button. When I let suspend happen by itself (I mean leaving the PC alone and using the idle timer set in the power settings, in my case 30 minutes) this results in black screen. Monitor led keeps blinking as in stand by mode and sometimes I get an on screen message from the monitor saying "no recognised video signal). Monitor and GPU are connected through HDMI. A hard reboot then is the only way to get back into my desktop.

[SMG]

1) The update mentioned was only for the Mint system through the usual Mint Update, so no proprietary nvidia drivers since those are offered through the Device driver manager (in Cinnamon).

Actually, once you have installed the Nvidia driver using Driver Manager, it is my understanding any updates to the driver are offered through Update Manager.

If you open Update Manager and select View > History of Updates, you will see what you have updated and when.

There was a kernel update, however.

A kernel update is an easy update to "undo". You can boot into grub, select the Advanced Options, and then select the prior kernel for booting. If you would like to try that and need directions, let us know.

Strange thing is: sometimes the computer resumes normally. This mostly when I manually go into suspend mode (menu > shutdown > suspend) and resume by pressing power button. When I let suspend happen by itself (I mean leaving the PC alone and using the idle timer set in the power settings, in my case 30 minutes) this results in black screen. Monitor led keeps blinking as in stand by mode and sometimes I get an on screen message from the monitor saying "no recognised video signal). Monitor and GPU are connected through HDMI. A hard reboot then is the only way to get back into my desktop.

Do you use a screensaver before it goes to suspend (when you let it happen by itself)? If so, maybe try turning the screensaver off and see if that makes a difference.

[sebbouckaert]

Actually, you are right about the way nvidia drivers automatically are updated. I saw that the most recent update of the nvidia driver to v. 460 happened in July. So this version 460 has been running since then until a few days ago without issues.

I will try booting with an older kernel version as you describe, and will see what happens next. Hope I didn't delete older kernels with apt-autoremove. Keep you posted.

Nope, not using any screensaver.

[sebbouckaert]

I actually found a USB drive with timeshift snapshots. Thing is, when I plug it in it keeps clicking away and does not mount. Oh well, Murphy's law I guess...

So now I'm stuck between a rock and a hard place: older nvidia drivers seem to work well with suspend and resume but give me horrible tearing issues in Firefox (and on the desktop) which would take several hours finding out (again) how to fix these. The newer 470 driver does not give any tearing issues but seems not to like the latest kernel 5.4.0-84-generic installed.

Two options, I guess:

1) booting in to GRUB and selecting an older kernel. Questions:

can someone point me to a step by step manual for this? I managed to press 'e' during boot and could access the options for booting 20.2, selecting an older kernel. However, I wonder how to make this selection permanent? This also boots in a very low screen resolution, since no according graphic kernel module is loaded and I would have to install a nvidia driver during first boot. But I hope the desktop is accessible since this ridiculously low resolution on my 4K monitor.
Then next, how would I prevent future kernel updates running an older one? I guess I could 'opt out' for kernel updates in the update manager somehow? Does running an older, unsupported kernel give me any security issues (like logging in to my bank etc.)?

2) Do a complete reinstall of 20.2 hoping everything will work (again) out of the box? As you all know, a lot of hassle, configuration etc.

So any help, specifically on solution 1) would be appreciated.

[SMG]

I will try booting with an older kernel version as you describe, and will see what happens next. Hope I didn't delete older kernels with apt-autoremove.

You can always re-install an older kernel.

Open Update Manager, select View > Linux Kernels and click Continue. In the left pane click 5.4 to see the available 5.4.0.x kernels. The middle column will let you know which one is active and which ones are installed. To install an older kernel, just click on the version number and the "Install" button will appear.

You can also remove installed kernels through that interface. You can not remove the kernel currently being used.

I actually found a USB drive with timeshift snapshots. Thing is, when I plug it in it keeps clicking away and does not mount.

As long as you have tried different ports with the same result, then I should probably make sure I have more than one usb with snapshots so that does not happen to me.

1) booting in to GRUB and selecting an older kernel. Questions:

can someone point me to a step by step manual for this? I managed to press 'e' during boot and could access the options for booting 20.2, selecting an older kernel. However, I wonder how to make this selection permanent?

After you boot with the older kernel, go into Update Manager like I described above and remove the newer kernel. Mint will always boot to the highest number kernel so if you make the older kernel the "highest number" by deleting the others then you will automatically boot to that older kernel.

This also boots in a very low screen resolution, since no according graphic kernel module is loaded and I would have to install a nvidia driver during first boot. But I hope the desktop is accessible since this ridiculously low resolution on my 4K monitor.

This sounds like you may be selecting a kernel with (recovery mode) after it. Select an option without that in the name because Recovery Mode disables graphics drivers.

Then next, how would I prevent future kernel updates running an older one? I guess I could 'opt out' for kernel updates in the update manager somehow?

Rather than opting out of future updates, I would recommend testing newer versions as they come out. If you have problems, you know how to can go back to what you were using. It's possible a future update may address whatever issue your system is experiencing.

Does running an older, unsupported kernel give me any security issues (like logging in to my bank etc.)?

It depends on what changes are in the kernel. Some are just bug fixes for drivers that were not working properly. I would think running an older browser would be more of an issue with the example you provided.

[sebbouckaert]

Thanks for taking time to help me out, most appreciated, SMG!

[SMG]

You're welcome.
I'm curious to see if the kernel change was related to this issue, so let us know when you get a chance to try it.

[sebbouckaert]

Just when I started to think how much a PITA Linux distros sometimes can be for a moderate user like me (as much as I love Mint)I had a breakthrough.

At first when I managed to boot kernel 5.4.0-81 (and make that permanent by deleting 5.4.0-84, as you said) the installed 470 driver would not load, and I was sure not to boot in recovery mode, so no further 'point-and-click' there...

But playing around with tty levels made it possible to install GPU driver v460 for this kernel from CMD. To make sure not to mess drivers around I first completely removed v470 and then reinstalled 460.

What happened next was 'funny': the screen resolution was back to normal, but now the computer did not want to suspend at all any more, instead kept showing the login screen after going into suspend mode.

But then I upgraded the driver once again to v.470 and rebooted and all seems well now, suspending and resuming fine.

So, as for now I can say that for my hardware combination running kernel 5.4.0-81 and nvidia driver 470 seems to have solved this issue. I'll keep trying suspending/resuming, of course. If any future issue persists I'll let you know.

[SMG]

So, as for now I can say that for my hardware combination running kernel 5.4.0-81 and nvidia driver 470 seems to have solved this issue.

Time for a Timeshift snapshot! It looks like you have plenty of room in your root partition to take and save one snapshot. You do not have to set up Timeshift to run automatically; just do a one-time snapshot. That way if you run into any future issue you will have it available.

[sebbouckaert]

Yes, found another HDD lying around here Formatted it and took a single Timeshift snapshot, hoping to make things easier with future issues.

[sebbouckaert]

Update: after the problem seemed to be fixed as mentioned above, there was a regression resulting in the same black screen again.

I got a little fed up and so went for a complete clean reinstall of Uma 20.2. After the new system was fully updated, I found myself once again in the same situation, with no changes in behaviour.

On top of that, there was another graphics issue now: drivers 460 and/or 470 would not load at boot, despite properly installed. I could solve this with adding

Code: Select all

nvidia 
nvidia-drm
nvidia-modeset
 

in /etc/modules, apparently making the modules load faster at boot as described here:

https://forums.linuxmint.com/viewtopic.php?t=331611

Hope this is helpful for others.

Then yesterday another kernel update was offered, this time from 5.4.0.84 to 5.4.0.97. Again this gave me the same black screen after resuming in combination with drivers 490 and/or 470.

Restoring a Timeshift snapshot with kernel 5.4.0.84 made sleep/resume work again.

This ongoing trouble with current nvidia drivers and kernels is pretty annoying since everything has been working fine for years/months. I am no expert so I don't know what changes 'under the hood' have happened recently. All I can hope is that things get solved (on both sides, kernel wise and driver wise) in future updates. Since I see many similar threads I suppose a bug about this has already been reported.

[SMG]

Then yesterday another kernel update was offered, this time from 5.4.0.84 to 5.4.0.97.

The update I received yesterday was from 5.4.0-84 to 5.4.0-86. I have no idea what the 5.4.0-97 might be.

This ongoing trouble with current nvidia drivers and kernels is pretty annoying since everything has been working fine for years/months. I am no expert so I don't know what changes 'under the hood' have happened recently. All I can hope is that things get solved (on both sides, kernel wise and driver wise) in future updates. Since I see many similar threads I suppose a bug about this has already been reported.

I did notice Nvidia released an updated driver on September 20th which mentions two different bug fixes that seem related to suspend and resume issues. It usually takes a little time for the package managers to prepare the Nvidia updates so they are available in Driver Manager, but I expect there will be a Nvidia 470.74 version available soon. I'm hoping this clears a lot of the issues people have been having.

[sebbouckaert]

You are of course absolutely right about kernel version 5.4.0-86. Same as earlier with quoting nvidia driver versions, I messed up numbers as I suffer from dyscalculia. This in combination with typing at a fast rate, my apologies...

[SMG]

No problem. Some people have installed kernels with what I consider to be "odd" numbers and those were the actual numbers of the kernel. I didn't know if maybe you found a new kernel source.

[sebbouckaert]

A thing I found in an similar thread elsewhere on this forum is that when i launch

Code: Select all

sudo pm-suspend

from the terminal

resuming seems to work fine. I don't know exactly what the difference is with using the power/suspend button from the main menu. I however notice a different behaviour. This command makes my system go to sleep immediately. The menu option gives me a sort of greenish screen first for a second or so. Also, when resuming it immediately shows the deskrop whereas the usual option first shows the login screen (if not a black screen in my case).
Anyway, seems like a reasonable workaround for now...