Replying to https://github.com/ValveSoftware/Proton/issues/7442#issuecomment-1909589488

i dont have vlk installed either and im running endeavour os with the latest packages installed, the only difference im seeing is that your running a wm while im running kde plasma 5.27.10 and i have tried both kwin x11 and kwin wayland with the same results of it crashing when entering the intro of creating a game

enshrouded.log
here is the enshrouded crash log aswell

thats odd the game is working on kde neon unstable with flatpak steam

edit: if i set reflections to quality the game will crash, setting everything to normal/balanced doesnt make the game crash and its working on endeavour os now so yay :)

On a fully-updated EndeavourOS, with the following stats (and skipping Steam's shader precaching):

GPU: AMD Radeon RX 6650 XT
CPU: AMD Ryzen 5 1600 x12
DE: Gnome 45.3
Kernel: Linux 6.7.1-arch1-1
GPU Driver:  amdgpu
Proton: 8.0-5

The game starts and appears to preprocess the shaders.

However, the problems start once the Keen logo shows up, followed by the main menu.

Under X11, when the window is not selected, the game is choppy, but appears correctly (which seems to be expected behavior from my memory of playing on Windows). When I mouse over the window, or click on the window to make it active, though, the window turns black. I can click a menu element (Play or Settings, for example) a frame or two before it goes black, but if I then make the window unselected to see the current state, it will appear back at the main menu. This is the same in either Windowed mode or Fullscreen mode (toggled with Alt+Enter).

Under Wayland instead, it's the same as X11 under windowed mode. However, if I switch in fullscreen mode with Alt+Enter, then the game wholly appears correctly, regardless of if the window is active or not. I still get a black screen game crash when attempting to load a world, though - about 6 seconds after the loading screen starts, my GPU hangs, first evidenced by my inability to scroll a web browser on my other screen, after which my mouse cursor also stalls.

About 12 seconds after that starts, the game crashes.

The crash log is as follows:

https://gist.github.com/SolarLune/205f98300b60908ad06f206999320bc6

I tried a few times more - setting the graphics quality settings to Max Performance didn't help; it still hung the GPU driver. Changing it from my desktop resolution (1440p) to 720p and turning off FSR2 as antialiasing didn't help either - it hung for a longer period of time this time, and then crashed Wayland (I presume) and kicked me out to the login prompt.

System:
  Kernel: 6.6.13-200.fc39.x86_64 arch: x86_64 bits: 64 compiler: gcc
    v: 2.40-13.fc39 Desktop: KDE Plasma v: 5.27.10 Distro: Fedora release 39
    (Thirty Nine)
CPU: AMD Ryzen 9 7950X3D 16-Core @ 32x 5.759GHz
Graphics:
  Device-1: AMD Navi 31 [Radeon RX 7900 XT/7900 XTX] vendor: Sapphire NITRO+
    driver: amdgpu v: kernel arch: RDNA-3 bus-ID: 03:00.0
  Display: wayland server: X.org v: 1.20.14 with: Xwayland v: 23.2.4
    compositor: kwin_wayland driver: X: loaded: amdgpu
    unloaded: fbdev,modesetting,radeon,vesa dri: radeonsi gpu: amdgpu
    resolution: 1: 2560x1440 2: 2560x1440 3: 2560x1080 4: 2560x1440
  API: EGL v: 1.5 drivers: radeonsi,swrast platforms:
    active: wayland,x11,surfaceless,device inactive: gbm
  API: OpenGL v: 4.6 compat-v: 4.5 vendor: amd mesa v: 23.3.3 glx-v: 1.4
    direct-render: yes renderer: AMD Radeon RX 7900 XTX (radeonsi navi31 LLVM
    17.0.6 DRM 3.54 6.6.13-200.fc39.x86_64)
  API: Vulkan v: 1.3.268 drivers: radv,llvmpipe surfaces: xcb,xlib,wayland
    devices: 2

I was able to start a single player game running Proton Experimental: Bleeding Edge.
I also had to turn off a particular graphical setting regarding some voxel rendering option in order to have normal frames, otherwise it would render less than 1 FPS.

I do still get an crash of the display driver / display environment if I attempt to join a dedicated server.

Replying to https://github.com/ValveSoftware/Proton/issues/7442#issuecomment-1910828896

That seems to be the same problem for me, hosting works tho

Made mesa issue for the hang when entering a world https://gitlab.freedesktop.org/mesa/mesa/-/issues/10500

Edit: reworded as it likely isn't the intro video or entering the world for the first time part that is the cause.

FWIW, I updated to a kernel-vanilla/stable 6.7.1 from Fedora copr, and the bad behavior changed.
I can now connect to dedicated servers and play mostly normally, and Voxel Detail Models can be enabled now without killing frames.

Point Light Shadows are now broken though, and cause really low frames if enabled.
Otherwise, runs great.

Worth noting that there were several hotfixes targeting the game today. I don't know if any of those would have had a direct effect on our issues.

Back with another finding. As per ProtonDB, starting the game and loading into a single-player world works if I have the RADV drivers installed. On EndeavourOS / Arch, this is done through installing vulkan-radeon and vulkan-icd-loader, I believe, and then running the game with the command VK_ICD_FILENAMES=/usr/share/vulkan/icd.d/radeon_icd.x86_64.json:/usr/share/vulkan/icd.d/radeon_icd.i686.json %command%.

Also note that the game runs fine on Steam Deck at about 25-30 FPS on all Performance settings.

I'm also having similar problems, the game launches fine but when I join a multiplayer game it starts to load then I get a black screen, then loading screen again and after that the whole system hangs and I have to pull the power to restart computer.

I have also found a way to circumvent this hang, if I start a single player game first and then quit to start screen I am able to join a multiplayer game without the system hanging. Doing it this way I loose about 2/3 of my fps though. going from 60-100 to 19-25, mangohud is telling me my GPUs memory frequency is going down from a locked 1000Mhz in the single player game to a fluctuating 400-600Mhz during multiplayer.

System:
Nobara 39
Mesa 23.3.3
Kernel 6.7.0
Proton, its the same behavior with proton 8, experimental and ge-proton-8-30

The game runs fine on Ubuntu 23.10 with a RTX 3080 and Ryzen 7 5700X, even at some times beating Windows 11 performance on the same machine.

However the only glitch I have found is that sometimes when launching the game, it will launch in windowed rather than full screen, but this can be easily fixed by toggling in game.

Whether this is the game, Proton or Gnome is anyone's guess.

I have similar issues. In order to be able to join a multiplayer dedicated server, i need to sett my settings to "Max Performance". When in game again, i turn them on one by one.

The following options causes issues for me:

• Contact Shadows: Crashes the game. Sometimes i get a crash message, sometimes i need to kill the computer.
• Reflections: Crashes the game. Sometimes i get a crash message, sometimes i need to kill the computer.
• Point Light: Kills frame rates. I drop to 2-3 fps on my 7900XTX

I can have all other settings on max without issues.

System:
Kubuntu on kernel 6.7.6
AMD 7800X3D
Sapphire RX 7900XTX Nitro+
64Gig 6000mhz memory

Edit:
It seems with those 4 settings disabled, I am able to launch the game as is. The strange thing is that sometimes i can enable some of the settings without it crashing. But not always. But i have steadily played with these off now for a two weeks and no more crashes.

Edit 2:
I can also reliably launch the game with Voxel detail models set to max now. It crashes very rarely with this setting on.
Still not possible to launch with the above 3 settings on anything than low/off.

The game feels very stuttery for me on Wayland as VRR is not being used even when forced to always on in Plasma settings. The game only has windowed or borderless mode and trying it in windowed mode then using KWin to fullscreen it (which enables VRR with every other program, you can tell by slight flickering) does not work for Enshrouded.

with the newest patch it crashes a couple of times when trying to load the shader cache and trying to load into a world will stay in the loading screen for a couple of minutes then it kicks you back out to the menu, also some images are broken

this was under latest proton experimental(bleeding edge)

rx 6700 mesa 24.0.3 kernel 6.8

EDIT: latest proton experimental(bleeding edge) the game is now working properly, able to load in and main menu looks correct!

Can confirm, the remaining issue is the never ending 'validation' of around 10GB. Any known fix for this?

Can confirm, the remaining issue is the never ending 'validation' of around 10GB. Any known fix for this?

Yeah I am getting the same issue. 9.2 GB of data is being downloaded every time. I have a feeling it is due to the shader cache.

EDIT: Yeah I left Steam to finish processing the shaders in the background and now I no longer get the 9.2 GB download.

EDIT 2: Nvm, it still seems to trigger big downloads but only after a day or so. Doesn't happen with every launch of Steam.

Another issue I am experiencing is if I set Texture settings to Max Quality, all the textures appear black. If I drop it down by one to Quality, it seems okay.

@kisak-valve Here's the proton logs: steam-1203620.log

To generate these logs, I booted into the game, set the texture quality from Quality to Max Quality to trigger the black texture bug, and then loaded into my world. After spinning the camera around for a few seconds, I immediately quit the game.

Here are my Steam Runtime Diagnostics

For anyone curious, this is what it looks like in game:

https://github.com/ValveSoftware/Proton/assets/241316/6a807e26-a56a-49ea-9d8e-3aec84a4c5e0

Another issue I am experiencing is if I set Texture settings to Max Quality, all the textures appear black. If I drop it down by one to Quality, it seems okay.

Thank you for reporting this issue. I was able to determine this is an NVIDIA driver regression only present in r555. A fix will be shipped in a future driver release. If you wish to reference this bug in the future it is being tracked internally with bug number 4675675.

with the latest enshrouded update the studio improved steamdeck support and added fsr3 which is working perfectly under proton experimental(bleeding edge)

the game also doesnt crash or freeze when switching around graphics settings so there are no more problems that i am seeing now, this under my rx 6700, mesa 24.1.4, kernel 6.10, cachy os

As of last week something broke when running the enshrouded gameserver in proton, should this be tracked here as well? See https://github.com/ValveSoftware/Proton/issues/7995 (it does not appear to be limited to Fedora) and the linked ticket to find some additional testing.

FWIW, I use ProtonGE 9.10-9.12 on my desktop, I use Nvidia 555 on Fedora 40 Workstation (GNOME) and the actual game works just fine.

Steam (native)
ProtonGE 9.12

but yes, the enshrouded dedicated server in a container with proton is broken.

I use kernel 6.6.3, mesa 23.3.0 and amd fury r9 card and the game crashes after 1 minute with proton 9.03, the system is frozen and reset is needed. I also used proton experimental, ge-20 and got even more crashes. Many steam games work very stable with my configuration. I even do not remember such dreadful crashes with them. They were maybe 5 years ago. Now if a game starts running it continues to work in my system. Thus, crashes are a huge problem of this game.

steam-1203620.log
My log and journal log https://gist.github.com/pwasiewi/bb087999f0e1a3c7ce75141a3150821c

@pwasiewi You said you used kernel 6.6.3, but in your logs it says you're on 6.11.3-zen1. There's a kernel bug introduced in 6.10 which breaks the zerovram workaround (as described here: https://gitlab.freedesktop.org/drm/amd/-/issues/3812), so that could be why it still crashed for you with RADV_DEBUG=zerovram.

I tested Mesa 24.2.7 (and later 24.3.1) on kernels before and after the regression with RX 6650 XT to make sure it wasn't a new type of hang. I got four hangs with the newer kernel in four hours total, and none with the older kernel in three hours total. My repro rate wasn't great, but I would've expected at least one hang in the time I tested on the older kernel if it was a new/separate issue, so it seems likely that this is the same issue as before.

Until there's a fix for the kernel bug, the game should be fine on a 6.6 kernel (with Mesa 24.0.8 or newer, or RADV_DEBUG=zerovram).

Hey thank you i will report the Problem to Mesa then.

One of the few games I couldn't get to work on the Radeon 9070 XT so far. Works fine with the Radeon 7800 XT with the same setup.

Kernel 6.13.5
Mesa 25.0.0
kernel-firmware-amdgpu-20250302-1.1 (openSUSE Tumbleweed package)

Crashes with this:

[W 00:00:00,574] [graphics] vkQueuePresentKHR failed with error 'VK_SUBOPTIMAL_KHR'
[W 00:00:00,585] [graphics] vkQueuePresentKHR failed with error 'VK_SUBOPTIMAL_KHR'
[W 00:00:00,602] [graphics] vkQueuePresentKHR failed with error 'VK_SUBOPTIMAL_KHR'
[W 00:00:00,618] [graphics] vkQueuePresentKHR failed with error 'VK_SUBOPTIMAL_KHR'
[I 00:00:00,642] [app] start creation step Preload
[W 00:00:00,684] [graphics] vkQueuePresentKHR failed with error 'VK_SUBOPTIMAL_KHR'
[E 00:00:00,696] [graphics] Failed to allocate download buffer memory.
[W 00:00:00,698] [graphics] vkQueuePresentKHR failed with error 'VK_SUBOPTIMAL_KHR'
[E 00:00:00,701] [graphics] Failed to allocate download buffer memory.
[W 00:00:00,705] [graphics] vkQueuePresentKHR failed with error 'VK_SUBOPTIMAL_KHR'
[I 00:00:00,902] [online] User stats received.
[E 00:00:10,863] [graphics] vkWaitForFences() timed out after 10.00 s.. probably a driver crash!
[E 00:00:10,863] [graphics] Vulkan device lost...
[I 00:00:10,863] [DeviceFaultEXT] A GPUVM fault has been detected
[I 00:00:10,864] [DeviceFaultEXT] Caused by 1 in address range [ 18446603339462164480, 18446603339462168575 ]
[E 00:00:10,864] Stopping process due to unrecoverable GPU crash!

Tried Proton Experimental and bleeding-edge.

@C0Nd3Mnd there is a merge request for this referenced here: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/34027

I tested it and it works with the 9070 XT, but for now you have to build mesa from source (not released or even merged to main yet)

The game stays locked on the Preprocessing screen. The log indicates a failure to allocate memory but my RAM has still plenty of space according to the task manager.

I tried to launch the game with Proton 9.0-4, Proton Experimental and GE-Proton9-27.

Device: AMD Radeon RX 590 Series (radeonsi, polaris10, ACO, DRM 3.59, 6.8.0-57-generic) (0x67df)
Kernel: 6.8.0-57-generic
Mesa: 25.0.3 - kisak-mesa PPA

enshrouded_2025-04-09T11-16-05.log

EDIT: Retried today, after some updates of the system. Works with Proton 9.0-4.

I have a different issue where the gpu hangs randomly.
It doesn't happens on every play, but only when loading a save. I have included a trace but umr logs are empty, probably because the gpu resets before it could take a trace.

Software information

Archlinux

System information

• GPU: Radeon RX 6700 XT
• Driver: Mesa 1.25.0.3 and GIT (the trace one)
• Proton version: experimental_be

Log files

steam-1203620.tar.gz

radv_dumps_1556046_2025.04.09_18.11.06.tar.gz

dmesg.log

Not sure if this is a proton issue, but noticed the fossilize replay taking ages for this game (also saw people mentioning in reviews that shader precaching was taking a long time).

Noticed that the shadercache file is enormous:

$ stat -c '%s' .local/share/Steam/steamapps/shadercache/1203620/fozpipelinesv6/steam_pipeline_cache.foz
13580667237

13.5Gb cache file.

$ fossilize-list .local/share/Steam/steamapps/shadercache/1203620/fozpipelinesv6/steam_pipeline_cache.foz --tag 4 --size|sort -n -k 2|tail -4
6037bf64a9fe6952 3335367 compressed bytes, 8025120 uncompressed bytes
5087898493216e29 3343291 compressed bytes, 8031903 uncompressed bytes
234b5e467d71bda2 3344574 compressed bytes, 8029806 uncompressed bytes
41146b3da0507261 3344596 compressed bytes, 8030639 uncompressed bytes

Seems it includes the entire shader hlsl source code a couple of times. Probably an issue with the game, but maybe something that could be more efficient in proton too or going wrong? Not too familiar with this stuff.

ie. snippit like this is present 14x times, in the disasm output

$ fossilize-disasm .local/share/Steam/steamapps/shadercache/1203620/fozpipelinesv6/steam_pipeline_cache.foz --filter-module 41146b3da0507261 --output /tmp/ --device-index 0

...
  47   │ OpSource HLSL 620 %21 "#line 1 \"E:/keen/checkouts/Default//holistic/content/shader/platform.hlsl\"
  48   │ 
  49   │ #ifndef KEEN_PLATFORM_HLSL_INCLUDED
  50   │ #define KEEN_PLATFORM_HLSL_INCLUDED
  51   │ 
  52   │ // rebuild counter: 10
  53   │ 
  54   │ #ifndef KEEN_CONCATENATE
  55   │ #   define KEEN_CONCATENATE_HELPER( value1, value2 ) value1 ## value2
  56   │ #   define KEEN_CONCATENATE( value1, value2 ) KEEN_CONCATENATE_HELPER( value1, value2 )
  57   │ #endif
  58   │ 
  59   │ #define KEEN_IDENTITY( value )              value
  60   │ #define KEEN_MAKE_VARNAME( prefix )         KEEN_CONCATENATE( prefix, KEEN_IDENTITY( __LINE__ ) )
  61   │ #define KEEN_STATIC_ASSERT( EXPRESSION )    typedef uint KEEN_MAKE_VARNAME( keen_static_assert ) [ (EXPRESSION) ? 1 : -1 ]
  62   │ 
  63   │ #if defined( KEEN_FLOAT16 )
  64   │ typedef float16_t float16;
  65   │ typedef float16_t2 float16_2;
  66   │ typedef float16_t3 float16_3;
  67   │ typedef float16_t4 float16_4;
  68   │ #endif
  69   │ 
  70   │ #if defined( KEEN_PLATFORM_PS5 )
  71   │ #   define KEEN_STRUCT_BUFFER           RegularBuffer
...

successfully connected on 2 servers with proton experimental-10.0-20250612c

regards

successfully connected on 2 servers with proton experimental-10.0-20250612c

regards

Thanks heaps! Can confirm working as expected now with Proton Experimental.