This bug makes me so tired: https://bugs.launchpad.net/ubuntu/+source/chromium-browser/+bug/1894627
I have a 970 and the desktop is a total mess after resume from suspend to RAM.
This instruction makes the machine hang on resume: Chapter 21. Configuring Power Management Support
Is there a fix planned? You can not seriously think that each and every one of the apps need to be rewritten because of https://www.khronos.org/registry/OpenGL/extensions/NV/NV_robustness_video_memory_purge.txt ?
1 Like
Can you please generate a bug report log after a failed resume with the vidmem preservation stuff enabled? If you can SSH into the system after the failed resume that would be best. Otherwise, generating a bug report log after a reboot would hopefully still provide some useful information about what went wrong.
It might also be helpful to attach the output of journalctl -b -1
to capture as much information about the failed resume as possible.
Thanks henrik4 for pointing out the almost 5 year old known issue with the Linux NVidia driver!
I’ve spent hours searching through multiple sleep/hibernation bug reports involving the Linux NVidia driver, and this is apparently not well known. I too would like to know if there is any expectation of a timeline for when “planned architectural evolutions in the NVIDIA Linux driver stack will allow video memory to be persistent” (to quote the khronos.org
link above).
Various NVidia drivers have issues on resuming from sleep for me, while running KDE on Ubuntu 20.04. Have tried numerous things to resolve, including different drivers, and other hacks being recommended on this and multiple other sites. If one looks hard enough, we can pretty easily come up with dozens of examples of such reports, all involving NVidia Linux drivers.
The one thing that seems to resolve this for me for the most part, is to kill applications that are using GPU memory before requesting sleep. I rarely have problems when I do this, but as shown in the output below, multiple applications required for KDE are always running in GPU:
Sun Jan 10 19:56:24 2021
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 390.141 Driver Version: 390.141 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 Quadro K2000M Off | 00000000:01:00.0 On | N/A |
| N/A 34C P8 N/A / N/A | 642MiB / 1999MiB | 0% Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 2798 G /usr/lib/xorg/Xorg 389MiB |
| 0 3105 G /usr/bin/kwin_x11 98MiB |
| 0 3113 G /usr/bin/plasmashell 113MiB |
| 0 5957 G /usr/lib/firefox/firefox 1MiB |
| 0 7099 G /usr/bin/systemsettings5 34MiB |
+-----------------------------------------------------------------------------+
I am not interested in providing logs to troubleshoot this, I only want to know that NVidia is working on this, and a general idea of when we can expect to see this feature fixed. Any ideas?
Like the OP already mentioned, the “planned architectural changes” are this:
https://download.nvidia.com/XFree86/Linux-x86_64/450.57/README/powermanagement.html
WFM, but seems to cause problems for other users.
For some reason, this problem went away. I reinstalled kubuntu 20.10 and then I got Nvidia drivers version 460. Now I don’t see the problem any more. I guess the new drivers have some improvements. Anyone know?
Or maybe something else randomly shifted…