Description of problem: Plasmashell process crash with core dump when restoring session from suspend with 570 Nvidia driver. Version-Release number of selected component (if applicable): dkms-nvidia-current-570.172.08-2.mga9.nonfree How reproducible: Always Steps to Reproduce: 1. Suspend your computer 2. Restore the session.
Created attachment 15051 [details] Journal of the crash
My system also experiences the same issue. Graphics fail to wake from sleep or hibernate mode after upgrade to 570.172.08. Must force a reboot to get system back. Operating System: Mageia 9 KDE Plasma Version: 5.27.10 KDE Frameworks Version: 5.114.0 Qt Version: 5.15.7 Kernel Version: 6.6.93-desktop-1.mga9 (64-bit) Graphics Platform: X11 Processors: 28 × Intel® Core™ i7-14700K Memory: 62.6 GiB of RAM Graphics Processor: NVIDIA GeForce RTX 4070 SUPER/PCIe/SSE2 Manufacturer: Gigabyte Technology Co., Ltd. Product Name: Z790 AORUS ELITE AX ICE
CC: (none) => leon244
Bug 34511 is probably linked.
CC: (none) => stephane.pontier
per doktor5000 at the Mageia forum (https://forums.mageia.org/en/viewtopic.php?f=23&t=15717&p=91658#p91658) I added the following to my /etc/modprobe.d/display-driver.conf file and the problem was fixed: options nvidia-drm fbdev=0 After doing this I was able to go into sleep mode and wake from it and go into X without a problem
(In reply to Leon Goldman from comment #4) > per doktor5000 at the Mageia forum > (https://forums.mageia.org/en/viewtopic.php?f=23&t=15717&p=91658#p91658) I > added the following to my /etc/modprobe.d/display-driver.conf file and the > problem was fixed: > > options nvidia-drm fbdev=0 > > After doing this I was able to go into sleep mode and wake from it and go > into X without a problem Thanks, it fixes this bug with suspend. However, does not fix all bugs coming from this driver, like Steam unable to launch games
Thank you for the reports Cyril and Leon, and doktor5000 for the workaround. I see you test this on mga9, but there are other reports on problems on Cauldron too. Assigning to maintainer --- I think I will add a note to errata9 about the problem and workaround, link to and describe it in https://wiki.mageia.org/en/Setup_the_graphical_server#Nvidia_proprietary_drivers
Assignee: bugsquad => kernelVersion: Cauldron => 9Whiteboard: (none) => MGA9TOOSee Also: (none) => https://bugs.mageia.org/show_bug.cgi?id=34511, https://bugs.mageia.org/show_bug.cgi?id=34512CC: (none) => friKeywords: (none) => FOR_ERRATA9
See Also: (none) => https://bugs.mageia.org/show_bug.cgi?id=34515
Generally, you users of Nvidia are *much* welcome to join QA to test Nvidia driver updates *before* we release! :-)
@Giuseppe: Maybe you can take a look at /etc/modprobe.d/display-driver.conf as at least the default values for nvidia-drm for fbdev option have changed (fbdev=1 was experimental before, is now the default), so our config might need to also be updated.
CC: (none) => doktor5000, ghibomgx
> @Giuseppe: > Maybe you can take a look at /etc/modprobe.d/display-driver.conf as at least > the default values for nvidia-drm for fbdev option have changed (fbdev=1 was > experimental before, is now the default), so our config might need to also > be updated. Yes, we can align the comment to the new default value, for instance like: # Disable the framebuffer device #options nvidia-drm fbdev=0 though for this bug it wouldn't have effect, since it's a comment. The default fbdev-> 1 has changed upstream since 560.35 or something near that. If using fbdev=0 is providing a smoother experience everywhere we might revert the default to 0, however what about Wayland? does it work with fbdev=0 and Wayland?
(In reply to Giuseppe Ghibò from comment #9) > > If using fbdev=0 is providing a smoother experience everywhere we might > revert the default to 0, however what about Wayland? does it work with > fbdev=0 and Wayland? This is about Mageia 9. Is Wayland an issue here?
CC: (none) => mhrambo3501
(In reply to Mike Rambo from comment #10) > (In reply to Giuseppe Ghibò from comment #9) > > > > If using fbdev=0 is providing a smoother experience everywhere we might > > revert the default to 0, however what about Wayland? does it work with > > fbdev=0 and Wayland? > > This is about Mageia 9. Is Wayland an issue here? No, it's working, but we need to be sure that if we change the default it doesn't cause issues with fbdev=0 under Wayland, as in the past upstream there was something related to requirement about fbdev=1 for Wayland.
I confirm that forcing the option nvidia-drm fbdev=0 by uncomment the line in /etc/modprobe.d/display-driver.conf fix the suspend bug, it does not fix Bug 34511 which seems related to Vulkan.
CC: leon244 => (none)
There is new version 570.172.08-2.mga9.nonfree, which uses fbdev=0 as default and should address latest problems.
(In reply to Giuseppe Ghibò from comment #13) > There is new version 570.172.08-2.mga9.nonfree, which uses fbdev=0 as > default and should address latest problems. Sorry new version is 570.172.08-3.mga9.nonfree in updates_testing.
(In reply to Giuseppe Ghibò from comment #14) I confirm that the revision 3 of 570.172.08 do not crash vkd3d games.(bug 34511). A bug that i forgot to mention, plymouth being weird : smaller size in the top left corner instead of fullscreen, is also gone. I had not experienced some of the bugs reported so i can't vouch for those but as far i could tell everything seem fine.
CC: (none) => dag42
(In reply to Didier Le Gloanic from comment #15) > (In reply to Giuseppe Ghibò from comment #14) > > I confirm that the revision 3 of 570.172.08 do not crash vkd3d games.(bug > 34511). > A bug that i forgot to mention, plymouth being weird : smaller size in the > top left corner instead of fullscreen, is also gone. I guess plymouth would probably need a refresh (at least in cauldron) to support that fbdev
Depends on: (none) => 34523
*** Bug 34515 has been marked as a duplicate of this bug. ***
Tested version 570.172.08-3. I cannot easily test the old Vivaldi but for me everything else that I have been having trouble with has been fixed by the new version. I have been running it for about 4 hours and have tested with steam and most of the common things I usually do, including suspend and resume. It all looks good to me at this point.
If I understand this correctly the restore from suspend per this bug description is fixed by Bug 34523 - Update request: nvidia-current-570.172.08-3.mga9.nonfree If you still experience that problem please reopen. Of other problems please open a fresh bug.
Status: NEW => RESOLVEDResolution: (none) => FIXED
Keywords: FOR_ERRATA9 => (none)