Unlike the Windows driver, the Linux driver drives my 10-bit display at 10bpc instead of 8bpc, which forces Display Stream Compression.
Can this be fixed please? Give us a configuration option like on Windows.
1 Like
Hi @abchauhan , to be clear does this issue you filed also encompass single GPU setups as well?
Iād filed an issue against Arch vulkan-tools package after finding both vulkaninfo and vkcube are no longer functional as of the 560 series release (had been confirmed working in 555, 550 series), but I later closed it:
$ vulkaninfo
'DISPLAY' environment variable not set... skipping surface info
wl_surface#6: error 0: DRM Syncobj surface object already created for surface 6
$ vkcube-wayland
Selected GPU 0: NVIDIA GeForce GTX 1050 Ti, type: DiscreteGpu
wl_surface#6: error 0: DRM Syncobj surface object already created for surface 6
Segmentation fault (core dumped)
Iāve also been following a seemingly related Mutter issue, but the developers point to this being strictly an nVidia driver problem:
Wayland clients break with Mesa - explicit sync related?
Current stack:
- nvidia 560.35.03
- kernel 6.10.10
- mutter 46.4
- mesa 24.2.2
- vulkan-tools 1.3.269
- vulkan-icd-loader 1.3.295
Thanks!
KDE 6.1.4. Wayland session. Open driver.
In Path of Exile ( vulkan renderer ), in rare occasions, the game just stuck on 1 frame and basically freezes. Most of the times, I can move mouse still/hear sounds from the game. In very rare occasions, my whole desktop freezes for few seconds, then i can again move mouse and terminate game.
Happens randomly, most of the times when other player join in my hideout ( loading mtx maybe? ).
Similar report from @adam228 and @steckums but from WoW.
Sep 09 14:30:28 Nix kwin_wayland[1874]: kwin_wayland_drm: Pageflip timed out! This is a kernel bug
Sep 09 14:30:33 Nix kwin_wayland[1874]: kwin_wayland_drm: Pageflip timed out! This is a kernel bug
Sep 09 14:30:38 Nix kwin_wayland[1874]: kwin_wayland_drm: Pageflip timed out! This is a kernel bug
Sep 09 14:30:43 Nix kwin_wayland[1874]: kwin_wayland_drm: Pageflip timed out! This is a kernel bug
Sep 09 14:30:48 Nix kwin_wayland[1874]: kwin_wayland_drm: Pageflip timed out! This is a kernel bug
Sep 09 14:30:53 Nix kwin_wayland[1874]: kwin_wayland_drm: Pageflip timed out! This is a kernel bug
Sep 09 14:30:53 Nix kwin_wayland[1874]: kwin_wayland_drm: atomic commit failed: Invalid argument
Sep 09 14:30:56 Nix kwin_wayland[1874]: kwin_scene_opengl: 0x502: GL_INVALID_OPERATION error generated.
Sep 09 14:30:56 Nix kwin_wayland[1874]: kwin_scene_opengl: 0x502: GL_INVALID_OPERATION error generated.
Sep 09 14:30:56 Nix kwin_wayland[1874]: kwin_scene_opengl: 0x502: GL_INVALID_OPERATION error generated.
Sep 09 14:30:56 Nix kwin_wayland[1874]: kwin_scene_opengl: 0x502: GL_INVALID_OPERATION error generated.
Sep 09 14:30:56 Nix kwin_wayland[1874]: kwin_scene_opengl: 0x502: GL_INVALID_OPERATION error generated.
Sep 09 14:30:56 Nix kwin_wayland[1874]: kwin_scene_opengl: 0x502: GL_INVALID_OPERATION error generated.
Sep 09 14:30:56 Nix kwin_wayland[1874]: kwin_scene_opengl: 0x502: GL_INVALID_OPERATION error generated.
Sep 09 14:30:56 Nix kwin_wayland[1874]: kwin_scene_opengl: 0x502: GL_INVALID_OPERATION error generated.
Sep 09 14:30:56 Nix kwin_wayland[1874]: kwin_scene_opengl: 0x502: GL_INVALID_OPERATION error generated.
nvidia-bug-report.log.gz (477.0 KB)
i get a lot of window tearing or incomplete window redraws when displays wake from suspend or even wake from screensaver. sometimes the screen saver blinks or flashes constantly.
Otherwise I have no other issues. Screen capture works with OBS now which it never did. Better Teams screen share experience too.
I am using arch kernel 6.10.10 with wayfire and a 1660ti
booting with these kernel parameters:
nvidia-drm.modeset=1 nvidia.NVreg_EnableGpuFirmware=0
Imo itās time for bug fixes release
Hi,
Thanks for reporting issue, could you please confirm how long you ran GravityMark Vulkan benchmark and observed nvidia stack trace.
Hi, can you specify the version that will contain the fix? I am on 560.35.03 and I still get the bug:
libnvidia-glcore.so.560.35.03!UnknownFunction(0xa10857)
libnvidia-glcore.so.560.35.03!UnknownFunction(0xe279e8)
libnvidia-glcore.so.560.35.03!UnknownFunction(0xe0d37f)
libnvidia-glcore.so.560.35.03!UnknownFunction(0xd4b78d)
libUnrealEditor-VulkanRHI.so!VulkanRHI::FFenceManager::WaitForFence(VulkanRHI::FFence*, unsigned long long) [/mnt/horde/++UE5/Sync/Engine/Source/./Runtime/VulkanRHI/Private/VulkanMemory.cpp:4690]
I am on Gnome wayland: GNOME Shell 46.5
4717891 as not been fixed yet i think. (Unreal engine crash after some time)
1 Like
Thanks for the info. Is there some public access to the bug tracker?
Itās pretty inconsistent. I run standard RT benchmark (fullscreen, 1920x1080, 200 000 asteroids), sometimes I get tones of those stack traces, sometimes none at all. Also they sometimes appear at random (rarely) without GravityMark running (but I launch other Vulkan applications from time to time, so it might be connected to Vulkan after all, for example, PCSX2, PPSSPP, RPCS3).
XID 109 error in games when tinkering with graphics settings a lot (in-game), like changing texture quality, DLSS profile, etc. + performance drop.
To reproduce:
- Open game (Steam / Proton Experimental / DXVK DX11 or VKD3D DX12 title)
- Start game
- Change graphics settings few times, from low to high/ultra and vice-versa.
- Game freeze with XID109.
OS: Arch (CachyOS) / Nobara (Fedora)
I think itās related to memory allocation.
No problems when i set graphics settings once before playing game.
Games affected (tested): any heavy titles like Alan Wake 2, Black Myth Wukong or The Casting of Frank Stone.
1 Like
Can verify the same issues with Warhammer 40K Space Marine 2 and Once Human. Its like the same AAA games released past 3 months have a common issue in VKD3D. Wukong as mentioned up in the post, had close to 50% performance loss on 4090 and 1440p. In 4k, it was close to not playable with Dx12 and RT.
Warhammer 40K Space Marine 2 is running around 80-106 fps max in CachyOs. Today I test the game in windows, where it starts at 160 fps in windows, where it would reflect maby Linux when its 90 fps. Something is broken delux somewhere.
Yeah, it seems to happen on windows with vkd3d-proton too. So itās sure some weirdness going on thatās not limited to the Linux driver only.
Nvidia again, please do whatever you got to do to fix the bottleneck
JavaFX applications randomly crash on newer drivers(including 560):
Current thread (0x000078921c1d20c0): JavaThread "QuantumRenderer-0" daemon [_thread_in_native, id=179506, stack(0x00007892137a7000,0x00007892137d3000) (176K)]
Stack: [0x00007892137a7000,0x00007892137d3000], sp=0x00007892137d0ce0, free space=167k
Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)
C [libnvidia-glcore.so.560.35.03+0x73f09c]
Java frames: (J=compiled Java code, j=interpreted, Vv=VM code)
J 4113 com.sun.prism.es2.GLContext.nBindFBO(JI)V javafx.graphics@23-ea (0 bytes) @ 0x00007892209df56c [0x00007892209df520+0x000000000000004c]
J 4559 c1 com.sun.prism.es2.GLContext.bindFBO(I)V javafx.graphics@23-ea (81 bytes) @ 0x0000789220a68434 [0x0000789220a683c0+0x0000000000000074]
J 4638 c1 com.sun.prism.es2.ES2Context.updateRenderTarget(Lcom/sun/prism/RenderTarget;Lcom/sun/javafx/sg/prism/NGCamera;Z)Lcom/sun/prism/impl/ps/BaseShaderContext$State; javafx.graphics@23-ea (266 bytes) @ 0x0000789220a82dec [0x0000789220a82d00+0x00000000000000ec]
J 3912 c1 com.sun.prism.impl.ps.BaseShaderContext.setRenderTarget(Lcom/sun/prism/RenderTarget;Lcom/sun/javafx/sg/prism/NGCamera;ZZ)V javafx.graphics@23-ea (278 bytes) @ 0x00007892209aa97c [0x00007892209aa740+0x000000000000023c]
J 3911 c1 com.sun.prism.impl.BaseContext.setRenderTarget(Lcom/sun/prism/impl/BaseGraphics;)V javafx.graphics@23-ea (47 bytes) @ 0x00007892209aa0ec [0x00007892209aa040+0x00000000000000ac]
J 4765 c1 com.sun.prism.impl.BaseGraphics.<init>(Lcom/sun/prism/impl/BaseContext;Lcom/sun/prism/RenderTarget;)V javafx.graphics@23-ea (155 bytes) @ 0x0000789220aacaa4 [0x0000789220aac6e0+0x00000000000003c4]
J 7224 c1 com.sun.prism.es2.ES2SwapChain.createGraphics()Lcom/sun/prism/es2/ES2Graphics; javafx.graphics@23-ea (272 bytes) @ 0x0000789220e25a5c [0x0000789220e25520+0x000000000000053c]
J 7104 c1 com.sun.javafx.tk.quantum.PresentingPainter.run()V javafx.graphics@23-ea (680 bytes) @ 0x0000789220df007c [0x0000789220def6c0+0x00000000000009bc]
J 4303 c1 java.util.concurrent.Executors$RunnableAdapter.call()Ljava/lang/Object; java.base@23-ea (14 bytes) @ 0x0000789220a16cfc [0x0000789220a16ca0+0x000000000000005c]
J 5014 c1 java.util.concurrent.FutureTask.runAndReset()Z java.base@23-ea (125 bytes) @ 0x0000789220afa7a4 [0x0000789220afa620+0x0000000000000184]
J 7218 c1 com.sun.javafx.tk.RenderJob.run()V javafx.graphics@23-ea (95 bytes) @ 0x0000789220e23614 [0x0000789220e235c0+0x0000000000000054]
J 7686% c1 java.util.concurrent.ThreadPoolExecutor.runWorker(Ljava/util/concurrent/ThreadPoolExecutor$Worker;)V java.base@23-ea (187 bytes) @ 0x0000789220ecabb4 [0x0000789220eca8a0+0x0000000000000314]
j java.util.concurrent.ThreadPoolExecutor$Worker.run()V+5 java.base@23-ea
j com.sun.javafx.tk.quantum.QuantumRenderer$PipelineRunnable.run()V+8 javafx.graphics@23-ea
j java.lang.Thread.runWith(Ljava/lang/Object;Ljava/lang/Runnable;)V+5 java.base@23-ea
j java.lang.Thread.run()V+19 java.base@23-ea
v ~StubRoutines::call_stub 0x0000789220400c86
siginfo: si_signo: 11 (SIGSEGV), si_code: 1 (SEGV_MAPERR), si_addr: 0x00000000000005ac
@amrits Any updates about frametimes/stuttering even with GSP off? Itās really frustrating to ask the same thing 5 times and get ignored.
Just wanted to chime in about the 560.35.03-2 driver. Iām using an RTX 2060 on Arch Linux with GNOME on Wayland, and Iāve been running into some serious freezing issues while playing Baldurās Gate 3 on Steam.
The game randomly freezesāsometimes right at the splash screen, sometimes while loading saves, or even during gameplay. I can still hear music and see the mouse pointer, but the game itself is completely unresponsive. Super frustrating!
I thought it might be a Proton issue at first, so I tried different Proton and Vulkan flags and even gamemoderun, but no luck. The only thing that worked for me was downgrading to version 555.58.02-1, which seems to have fixed the problem.
Has anyone else been dealing with this? Would love to hear if youāve found any other solutions!
I am also having similar issues as @fedev on a HP ENVY m7 Notebook that uses Pop_OS! 22.04, GNOME 42.9, kernel 6.9.3-76060903-generic, NVIDIA GeForce 940MX after being told to reboot to apply the 560.35.03 driver update, but on World of Warcraft, Lutris. After like 10 minutes of gameplay, if something with complex visuals triggers, like any or most spells, the game will freeze and crash, forcing me to kill it with a terminal. At first I thought it could be the wine runner (tried UMU-Proton 9 and the latest TKG bleeding edge version as of today - 917), but was told by that itās most likely the Nvidia driver, compatibility with flatpak, or something regarding its installationā¦ :( For the horde
Haha, welcome to the club. The best you can hope for is a vague acknowledgment of the issue. Letās hope that they are really cooking for the 565 betaā¦
Although I have not had stuttering issues with GSP off. I am hoping for multi monitor VRR finallyā¦
1 Like
HDR on my RTX 4070 on both gnome 46/47 and KDE 6.1.* and 6.2 is all washed out
HDR is washed out When enabled in Gnome 47 (#3681) Ā· Issues Ā· GNOME / mutter Ā· GitLab here is a report on the side of gnome.
but here is the kicker HDR used to work fine and it flickers in and out
Iām hoping for an acknowledgement of this HDR/WCG freeze issue. I havenāt seen anything from Nvidia about this. Iām still unsure if itās a driver issue or a KDE issue. If someone needs any logs or information to confirm this issue please just let me know whatās needed
1 Like